 So, I think we're started. Right, so, so welcome everybody. We graduate student summer seminar here. A couple of quick things before we get started. Just want to remind everyone that despite, you know, this being called a graduate students seminar we have people from, you know, various different backgrounds. So I just like to ask people to be respectful of everyone else. So please, please do ask questions and the spirit of a usual seminar it's perhaps easiest if you just go ahead and if you have a question just unmute yourself and slightly interrupt. I think that's totally fine. And if you if you don't want to say it out loud you can always just type it in chat I'll try to keep an eye on that and relay things to the speaker as they come up. I have a question that might involve sort of a longer more detailed answer. Perhaps easiest to hold it off until the end of a couple of minutes and you can ask them. And I think that is about it. So let me. This is our great, great pleasure to have Neil Coleman with us, who will be talking today about transitioning from academia to industry. So turn it over to you, Neil. Thank you. Thank you for the, the invitation. Let me get my slides shared. If you haven't seen them there in the, in the zoom chat but otherwise, I'm also going to throw them up on the screen here. We'll do the usual zoom song and dance can you all see my screen is it looking good. All right, let me see if, if I start the presentation, can you see the presentation or are you seeing the presentation mode. I am seeing just the slides, or just the side. All right, okay, we'll just, we'll just stick with the slide window here. So you get to see my presenter notes as well. All right. So I'm going to start with a little outline of what I'm going to talk about here. The first section is kind of my story background on me. What kind of math I like where I went to school and the kind of the decision process that led me out of academia into industry. I'll talk through my, my observations of academia and industry, kind of comparing contrasting the cultures, things that are good things that are bad in each one, and just share some of my observations and my thoughts about about the distinction having had a foot in both worlds. Lastly, I'll focus on skills that are transferable from academia from a from the PhD training that that one gets in graduate school, and which of those skills transfer over into industry, which which of those skills. And maybe don't transfer into industry, and then ways that you can frame your skill set to speak the language language that is more common in industry. The bias here will be kind of the white color tech industry because that's what I transitioned into. I have experience in other industries. So I can't really speak too much to them. So, just be aware that, you know, if you transition to say if you were thinking about transitioning to I don't aerospace design from from mathematics or aerospace research or communications research that there would be a different experience there but that'll be one of the themes that that we have in this talk. So let's jump in. Are there any initial questions. All right. Okay, so my story. I'll talk a little bit about me, who I am, where I come from. A little bit about the math that I like this is a math seminar so I feel like I should throw some kind of some kind of mathematical theme in my career so far. When I transitioned into industry when I've been doing since then. And then I'll talk through kind of some of the various decisions that I made in the considerations that affected my decision to, to not pursue an academic job. So, my academic resume is that I studied math physics and economics as an undergrad at Ball State University in Indiana, if any of you have heard of it. And then I went to Indiana University for graduate school. So I started in 2010. Indiana has required core math program so I taught each year for for undergrads. And of course I also could tutor on the side. I've transitioned out of academia in 2015 to 16. So that's when I was job hunting giving give a few job talks, settled to interviewing outside of academia and landed a job at all state, the insurance company. And then I completed my dissertation and I defended it the next year. So these days, I'm a data scientist or engineer, I kind of live at the intersection of analytics and statistics and software engineering. Previously I was so I work currently at a startup called Dina will share a little bit more about what we do in a little in a couple slides. And before then I was a predictive modeler or a data scientist at all state. Again, that's the auto insurance company. Kind of my personal life. So I live near Chicago. If you were here a little bit earlier I talked about Northwestern and the students coming back so kind of one of the northern suburbs of Chicago. I have a couple kids cat and a dog. You know the nice, nice quiet life here and some of the hobbies that I like to do I volunteer at the kids school so I'm like the kids chess coach. I like to play strategy games and I do enjoy reading science fiction. If any of you haven't checked out our little science fiction bookshelf which I realized after I was setting up is behind my right shoulder. The math I like so in grad school I specialized in spectral geometry, global harmonic analysis so this is that that famous paper by cats can one hear the shape of a drum. And the answer is, you know, if it's a 16 dimensional tourist drum. No, that's I think due to Milner. But if it's a convex, if it's a convex domain the answer is actually unknown. It was a result of Gordon and Web and Walpert in the early 90s that that in fact even planar domains one can there's there's domains you can get by gluing triangles together that that actually have the same spectrum. So this is the spectrum of the Laplace operator on on a domain, you fix boundary conditions and kind of crank the functional analysis. Machine and you get out the fact that it's compactly resolved so it has a discrete set of eigenvalues of finite multiplicity that converge at infinity. And this sequence of numbers, these eigenvalues is related to the geometry of the domain. So, for example, the asymptotic expansion of that sequence right it's very difficult to compute an individual one but kind of averaging it out in the, the large scale when can see that the its asymptotic growth is proportional to a power function whose coefficient is the volume of the domain. And then, you know, from there you asked start to ask other questions like well if, if they're iso spectral that is they have the same sequence of eigenvalues and multiplicities. Are they isometric answer, as I just said is no. There's other related questions like, can one drum the shape of a here that is given a sequence of eigenvalues can you find a Ramonian manifold whose sequence is that whose eigenvalues is that for any finite sequence I believe the answer is yet this is a yes. This is a result of Colin de Verriere. My dissertation in particular I thought about a variety of questions before I settled on. If you have two domains. You're playing with words can two drums the shape of a here so I've if I have two domains and I know something about their Laplace spectrum so I know that the case eigenvalue, ordered with multiplicity is strictly less of one than the other. So, to deduce about their comparative geometries. You know it is a dissertation so it's poorly written and I don't recommend you look it up it's embarrassing. The answer is a little bit not too much, but enough there to write and defend. During graduate school I was adjacent to the low dimensional topology geometry group. So, hyperbolic geometry planar geometry tecumular things modular spaces, and, and of course, I needed to learn some harmonic analysis as well so classical PDEs heat traces which kind of the trace of the heat operator the of the heat kernel, the trace of the wave kernel these are related to and give important techniques for understanding the spectral geometry or the global harmonic analysis of a domain. So this is kind of trying to establish a little bit I do know a little bit of math. I did go through grad school, even though that's not what this talk is about, you know, maybe maybe sometime I'll give a talk about about some of this math, but I'll pause here if there's, there's any mathematical questions about about what I've described here. All right, cool. So, 2015 kind of this is the highlights of my of my career so far 2015 I went on a job hunt started looking outside of academia landed a data science job at all state insurance companies you know very junior level data scientist, working on data science data to build insurance loss model so an insurance loss model says, if you have an insurance policy, based on what I know about you, how much should you be paying so that your expected loss, the expected pay out on your accident the accidents that you cause is covered by the present value of your payments. So, the question was so one of the big things for insurance companies is always getting more data sources better data sources to really improve the accuracy of those price models, so that if somebody is truly very risky, and you rate them as extra risky but you have data that your competitors don't. Your competitors say, Oh, you're just fine your average and then you turn out to be more risky and you cause them to have the other insurance companies to have more losses, which means they have to raise the rates on everybody which means more of those people will come over and get more fair prices from from your insurance company so that's that's the. That's kind of what I was thinking about using telematics data specifically so, for example if you have an app on your phone with a motion sensor in it, and we can detect if you're say, driving down the freeway at 90 miles an hour and swerving in and out of traffic turns out are more likely to cause accidents, surprise. So, working on ways to bring that data in and use it to really to make sure that that people pay their fair share. 2017 the next year I kind of spun off into erity which was a wholly on subsidiary of all state focused on on the collection and processing of telematics data and then turning that into data science products to sell. 2018 I transitioned to Dina my current my current employer. My role really expanded managing a lot of analytics data warehouse product data requirements I was the the 18th employee in the company's entire history. So there was a lot there to pick up a lot to be responsible for kind of working on that many different fronts for a couple years 2020 of course you know, we all had a year. And 2021 we're scaling growing kind of continuing to build our our product products out at Dina. My experiences at all state and Erity great experience, big corporations focused on really developing new people people without a lot of business experience before from kind of the junior level up to the senior level. And a great experience at Dina as well it's a very different experience. Small effective team, lots of hustling to to close deals to to really support individual customer initiatives. Kind of identifying and prioritizing you know, these five things can be fixed next month that this has to be done now, and a huge focus on building a strong product based on understanding patients and providers and what they need. A little bit about my current job, Dina. The focus of our company is helping people age in their homes, where a product company selling a software as a service products so we build software applications, big engineering team, big product team, of course lots of sales people this year as well. So really focus on expanding nationwide. We offer transitions from inpatient care settings to the home collaborative management of episodes of care. You know if you get a new hit at the hospital. And the hospital wants to make sure that you get the appropriate level of care and appropriate collaboration if your condition declines, you know if you're wound from the surgery gets infected. The hospital should know about that so they can help direct care to you that you need. Currently a lot of health systems. There's no collaboration. If you get a new hip health system will fax you out to fax out a bunch of referrals and you'll be accepted at home health, say a nurse shows up every week to your home or skilled nursing or rehab facility, and they won't talk after that. And then the hospital is on the hook often for paying for any follow up care that you need so it's good to make sure that that you get the appropriate level of care so we have tools that enable this. We also have tools that enable outreach so for example if someone is diagnosed with COVID. You may the hospital may say you're not severe right now we're going to send you home with a pulse oximeter, and then we provide a tool that allows the hospital to send a text message every morning saying hey, what is your pulse oximeter reading. And then if it drops below 95 or drops below 90, then a nurse gets an alert and can then follow up and schedule and we've we've actually had it's been really important and helpful. For example we've we've saved. People's lives according to at least according to testimony from some of our customers so that's been been really helpful and we've been focused on developing that this last year. And lastly, kind of our data product our insights. These are data transfers reports manage even that we provide to customers up to management by exception so at the start of a. Say, again, you have a hip replacement at the start of your home health care with the nurse visiting the nurse will fill out a panel of questions for you about your status and if something rings a bell, or is statistically associated with a decline. In condition we want to be able to take that assessment data and alert the nurse and alert the other people who are collaborating on your care to make sure that you get the right amount of care. So that that's a little bit about our product line. I don't mean this to be too much of an ad, but you know if you do have a connection, of course, reach out. We are a very small startup. We just, I wear many hats here so I'm the only data person I manage our data warehouse which is an archive of all the data that we've seen in the past. Today, I'm our data engineer. So I write programs that get data from other places into our data warehouse where it's archived and stored appropriately. I do data modeling. So I make sure that we're not missing data that we need to store that we're measuring the right quantities in a useful way. And I do statistical modeling to measure effectiveness and forget predictiveness of input data to identify opportunities return on investment effectiveness of treatments and so forth. And I've also done software engineering as well to build an application that production arises and operationalize is some of that those statistical models. I've got a link to our website here if you want to go check us out. We are also hiring, you know, if you're especially if you're a good fit for a data engineer role. As a senior data engineer role, let me know or if you know someone who is please feel free to pass on our information. I'll pause here. Are there any questions about what I do, or what I've done. Cool. So, a little bit about the modeling aspect. Yes, so when when you guys are say predicting things what what kind of things show up as as things you're predicting. So, so one thing, one example that we predict is re hospitalization. So, again, to my my example if you get a new hip at a hospital so a total joint replacement. And maybe you're on Medicare and Medicare pays for it. You would, if you are re hospitalized. That's a bad thing, right, it's bad because you had to go back to the hospital for whatever reason. It's bad because you didn't get the appropriate level of care in in rehab, or the hospital made a mistake. And for from the hospital's perspective, it's often bad as well, because Medicare will make them pay for it. So kind of all the incentives here are aligned to figure out who's going to re hospitalized be, you know, go back to the hospital, and who's not. And so one of the things that I do is, in fact, one of my current projects is going through a bunch of clinical assessment data, and identifying, you know, at time of discharge from hospital. Here's their data, they're the the clinical assessment. And how will they, what will happen to them, where they re hospitalized or not, are any of these extra predictive. So the process here is, you know, there's a literature review, a literature search to kind of because this is an interesting research question as well. And then a matter of seeing which of the literature's results can we reproduce in our data. And so it's hard to, excuse me, replicate right reproduction is taking someone else's data, and making sure that their experiment worked and their analysis worked replication is taking someone else's setup, gathering your own data, and making sure that that you are able to get the same results so replicating their results on on our data from from the various customers that we work with and in our data warehouse. So statistical modeling. It depends often, not to get too far in the weeds, but often we're interested in interpretability, because we're not replacing nurses we're augmenting we're helping clinicians with workflows to bring their expertise to bear on exceptions and so the clinicians need to know, they need to know why an alert triggered and why, you know, why do we think that this patient is at extra risk is it for reasons the nurses already internalized. Is it for reasons that the nurse was not yet not yet aware of and so understanding clinical workflows and understanding understanding the data that goes into the model is critically important to making sure that it's actionable and useful. Thank you so much, Bill. Did I did I get your question there. Yeah, yeah, thank you. You're welcome. All right. So, 2015. I'm a grad student in entering, ending my fifth year. The renewal letter was was much nicer from our department our grad chair, but it was effectively it's time to move out and get a job. 20 fall 2015. I started started looking. I gave a few job talks kind of sussed out some opportunities for math jobs. And I also thought through kind of what what would the different academic opportunities. So there's a research track where I get a postdoc somewhere in the world. And then they get another postdoc. And then I hopefully land a tenure track research job, do a lot of publishing and some teaching and some service. No job security for five, eight years until that the tenure packet goes in and then tenure, right the dream. There's also lots of teaching opportunities as well. One of the experiences that I really didn't appreciate as a graduate student who was doing a lot of teaching was gatekeeping for other departments that was the business school at Indiana University. One of the required for students who weren't directed myths to get at least to be in the classes that I taught. And just me teaching right these were big 80 person sections out of a class of maybe two or 3000 people each semester. And I didn't think that was a great. I got bitter about it. I'll be be honest that it was not conducive to students learning mathematics appreciating the beauty of mathematics. And I didn't appreciate being put in a position where, you know, I'm ultimately, you know, as much as we build process around it ultimately it's my decision. You know, which students pass the course which students fail and I didn't appreciate kind of being put in a position to to gatekeep a lot of their other opportunities that I weren't related. In my opinion to whether they were had really deeply absorbed the mathematics in the course. And then of course it's often in these positions difficult to have work life balance right to paraphrase the old engraving eight hours of work eight hours of grading eight hours of sleep. That's a little bit of an exaggeration, but the things that were important to me were stability. So at the time I had two young children. Now I have two medium sized children. So I wanted to put down roots in a community. I wanted to have good work life balance, you know, come home, be able to do things like volunteer at the kids school have family time with them. And I realized as I introspected what was important about mathematics was to me, and that I really enjoyed about it was kind of an intellectual playground, a series of abstractions and concepts that I could think through and put together and understand. But the the spectral geometry that the harmonic analysis that I did as much as as it was a as beautiful as it is as much as I enjoy it. That topic itself. That's that subfield was not critically important to me. And then there were other considerations, you know, having building a professional network. There were other people at Indiana University, who had transitioned to academia. Of course, salary is a concern consideration providing for for a family. And I wanted to find a place where, you know, even if one job didn't work out, there would be others. You know, Chicago, the Chicago area is a big metropolis, the second biggest on this whole continent, right the third on the continent, it's, there's a lot of job opportunities here, as opposed to a 10 year track or a postdoc. If something doesn't work out. What else is there, right? It's a small market. So the these decision points and these considerations kind of really led me to focus on finding a job in industry, not an academia. Now, and to be to be clear, in this presentation, I, I'm presenting my perspective. And I'm presenting my experiences and I, I don't want you to take this as a recipe or trying to persuade you away from a passion. I think it's, it's important just that everybody make a considered decision. And this was the these were the considerations in my decision. All right, that was a lot about me. We're at about the halfway point here. So I think we're more or less on target. The next up is next up I'm going to talk about some of my observations, having worked in academia as a graduate student and having worked in industry as as an individual contributor. I'm going to give some descriptions of the academic job market, the non academic job market, what career lighters often look like. And so I'm going to give you a little bit of an outline contrast kind of three interesting aspects of of the work and the work cultures, which are the pace of work, the freedom of work and power dynamics in the workplace. Academic career tracks. I've talked through tenure track jobs, research and teaching. There's also I want to call out there's also a large contingent faculty track as well and I mean it's, it feels like a dead end, but it is. And I don't know that career is necessarily the right way of talking about it. But it is often a stream of work that that people take on or are shunted into. And if I were to choose to stay in academia. This would be an option for the type of work that I would end up doing or that that one an academic can end up doing right semester to semester contracting paper class. Kind of a second, second class within a department's faculty sometimes depending on the department. You know, not necessarily part of department decision making, not necessarily getting benefits. You know, socially, socially precarious not not integrated into the core of the department, the tenure track and and tenured faculty industry, on the other hand, and speaking from my white collar tech perspective. You know, there's a variety of of roles that one can go into an adopt and there's fluidity across roles as well people can make career changes from engineering, one type of engineering to another from it to engineering to user experience to research and design and development. These are this is kind of a quick summary there's there's engineering which is focused on designing and building and maintaining systems. So there's software engineering but I also want to call out in. There's also mechanical engineering electrical engineering for the in the, the non software realm I don't have a lot of experience with that. Most companies these days almost all of them would have an IT department focused on managing technology for across the enterprise, as well as particularly the the area that I most familiar with shaping and storing data, building reports generating value from that stored data for the rest of the the company. There's user experience design so individuals whose whose whole focus is graphic design optimizing the experience of people as they interact with with the company, whether through an application interface, or through advertising. There's also technical research and design so identifying potential product opportunities, doing even more baseline research, performing experiments to confirm feeding into engineering and designs to build. There's also I'm going to call it two other potential tracks from the bottom up there's contract work so people can build careers as contractors. And generally contractors are brought on similar to how you might hire a contractor for fixing a building a new addition to a house, for example, or coming in and doing a single job taking a tree down fixing, fixing plumbing in the bathroom. A contractor is brought on to perform a single job, you know stand up a data warehouse, right. Let me report on this particular thing, do an analysis that predicts this particular outcome. So contract work is out there often it doesn't come with benefits but it's often more lucrative on an hourly basis than, than kind of the technical white color track where one embeds inside of an organization and performs a role within the organization. Although sometimes contractors are brought in as this this is my opinion is this is an unhealthy organizational design but sometimes a company will bring in contractors and embed them on teams and expect them to perform almost as as the, the full time employees, but this then creates a kind of a tiered class system within the workplace that I don't, I don't believe is healthy. And lastly there's management, which is recognized generally as a separate career track from the technical tracks or these individual contributor tracks. So it's a different set of skills. It's kind of functioning as connective tissue within the organization determining priorities, trying to identify what's the best way to organize the different roles and the different teams within within the company to produce planning out projects, executing projects, helping work proceed and making sure that people are talking to the right other people. You know knocking down people's doors, you know, if our, if our VP of engineering needs to review a design, and as a week overdue, the project manager will bug him until he does it. Right. So, different flavors there's people managers whose job is to make sure that that people feel, feel safe, feel heard have opportunities to grow their careers. There's product managers whose role is to make decisions and priorities about what a product should be. There's project managers whose role is to really lay out all of the individual tasks and make sure that everything is sequenced right, happening on time. So management I think is is and should be considered a different set of skills than individual contributing. And we'll revisit that theme here in a couple of minutes. This is a lot. I know I'm, I'm talking a mile a minute and my slides are mostly text so I'm going to pause here again for some more questions. Well, let's keep going then. So, just to get a sense of the job markets here. I went on math jobs this morning. And I looked at how many jobs were listed in the US. And here they are, there's 266. These are the website Glassdoor, and there are 7193 listed on Glassdoor, but these are not just any jobs these are the full time jobs that were posted last week. So I want to be clear. The scale, we're talking many orders of magnitude. This is everything that's posted currently on math jobs versus all the stuff that was just posted in the last seven days on Glassdoor 7193. So this is something to keep in mind and I think it's, it was an adjustment for me coming out of academia where you're as one specializes you're integrating into a very specific research conversation. Right. I could list out on on, you know, there's there's probably 10 or 15 people in the world who are in in global harmonic analysis. I know them. Even if they don't know me they know know my advisor. I'm contributing and it's a very small village and any contribution that I make in the in my field is is a contribution to a conversation with maybe 1520 people. Right. That's that scale the scale where there's 266 jobs in the market. Some are postdocs they're probably all scattered across all the different fields of mathematics to 7200 jobs listed in the last week alone on just one job posting site right there's many other job posting sites and companies won't use this. It's just a completely different level of scale and I think that's important to bear in mind when thinking about academia versus industry. Also notice the salary range. Of course, there's a lot of variation in salary right 55,000 in Chicago is very different from 55 in San Francisco. But notice that 55 is at the bottom of this range 55 up to 150. Of course, this is at all levels right from junior to senior. But even starting out it's it's comparing to it's comparing well to faculty tenure track jobs. Okay, so having, having seen the scale of the difference here between the two between the two professions here, I'm going to talk through kind of three cultural differences that I've noticed. The first that I want to address is the pace of work. So, in academia projects can take months or years right. How long did I spend writing my dissertation I don't even want to think about it, but it was it was quite a while. In industry, it depends heavily on the organization that that you're working for the smaller the organization, the shorter the time horizon. So, most of the work that I pick up. The projects are measured in hours or days. The longest, the longest project that I worked on that, frankly, should not have taken that long was a couple of months. So, the largest organizations will have the the capital to support and the time horizon to support deep research. And those organizations projects can take months or years. So, for example, if you're doing research for pharmaceutical organization, maybe you're developing an mRNA vaccine. Right, that that level of basic research can take years right they worked on the mRNA vaccination since the 80s. I think from when it was published, and often in those large organizations with those research labs there's a lot of collaboration with academia and academics in academia collaboration does go back and forth. Often it's very little structured often it's especially mathematics, it's with the, the individuals, you know, other people in your field in your subfield. It's very informal, very, very up to negotiation individual negotiation not not a formal negotiation but like hey how do you want to, you know, should we meet every week okay yeah that sounds good I'll send out a zoom right. In industry, however, the whole purpose of a firm is to collaborate. The whole purpose of a firm is so that we can is that a group of people can come together and work very closely together intensely to build something and to ship it out and create value. So there's a lot of communication, it's much more structured, especially in these these white collar in the tech industry there's a lot of a lot of structure around process to make sure that we're not wasting each other's time. We're communicating the right things to the right people. So there's a lot of communication between to manage expectations. It's it's very structured it's it's very different from collaboration in the academic sense. And lastly, the amount that you should think before asking questions. math. If you have a question for your advisor, maybe you go and think about it for a week and then you come back and you ask it and you think about it for another week. And then you add or maybe you think about it for a month and then you ask. There's often, you know, a set of status signals. People not not wanting to ask a question in seminar because it makes them look dumb. Right, there's asking a dumb question can maybe feels like a status demotion. Whereas in academia in industry, excuse me. And then ask the pace of delivery is so much faster shouldn't think more than an hour or two. Right. Although I'm sure the culture will vary considerably between organizations. But on that note, organizations in industry organizations are often very self aware about their culture and what they're trying to create. And I don't think that academia in academia. There's a lot of reflection about the culture that is being created and shepherded in a department. There's a story that I that I heard once Facebook was staffing up with a bunch of grad students or people with PhDs right out of graduate school. And a manager said to the team, you know, hey, if you don't get it right away, don't ask me think about it first and then come ask me if you've thought about it. The manager, nobody asked the manager any questions for the next week. And after the next team meeting the manager says, I told you guys to ask me after after you thought about it. But turned out the manager meant go think about it for 30 minutes and then ask me. But the grad students heard, think about it for a month and then ask me. So it's that that level setting on on pace that that it's very different. That was one of the big things that I, I frankly I struggled with coming out of academia. Freedom. There's a lot of freedom in academia right. That's the dream for research freedom. No formal constraints on your research direction, although it's driven by fashion. If you're writing stuff that nobody cares about. You're not going to get citations you're not going to get published. It's a tiny market in the sense that you own your own research you own it as a product. That in itself constraints, constraints your research. It's also a tiny job market as we discussed right 200 300 jobs in the whole US for for math jobs on industry. On the other hand, the focus is on providing value. So the, the company has customers. The focus of the company is to provide value for those customers so they continue to buy your products. So there's many formal constraints on research direction on the direction of your thoughts. Now, I mean, it's not thought police, but the things that you spend your time thinking about and doing are heavily constrained because often it's because it is driven by perceived customer needs. The difference. This is very similar actually to mathematics because, you know, there's no formal constraints on your research. But what you spend your time thinking about is driven by the needs needs of your customers. In other words, the people who will be reading and potentially setting your papers right you want to find a problem. This is a fundamental skill right to find a problem that is solvable and also interesting. The difference is that this is an internal constraint right you have to be aware of your market you're performing the function of a whole product team and also the all of the technicians on that product team. Whereas in industry, because it's a group of people collaborating to produce things of value. This has to be the these controls are controls are external. When I say controls, I don't mean, you know, formal control it's more being able to defend how you spent your time and defend that it was that it was based on customer needs. And lastly, in academia as we know work like balance can be difficult. Whereas in industry, there's a huge variation in work like balance from organization organization some you know you read about Netflix. They chew up their, their teams and spit them out right that they're very intentional about creating a culture where people are willing to prioritize work over everything else. So the company I work for Dina, you know, we have a lot of work and we have, it's very urgent work but it's, it's very focused on creating a healthy workplace for people to work and also have families and lives. And that's, that's a very intentional decision by our leadership as well. So again to organizations being self aware about their, their culture. And I should note that these constraints on research direction on as you grow in in a career and become more aware of the big product in your big picture in the organization and the product and the market and build trust with stakeholders. So the people that consume the output of your work. Are trusted to become more autonomous and kind of the mode of communication with the rest of the organizations which is from defending things that you think about to identifying the value you created and marketing it. So I do want to call that out that this is kind of from the junior to the senior contribution level, there will be a similar journey. I would imagine in academia as well. Lastly, I want to talk about power dynamics. I have a lot to say about this. So I'm going to be relatively brief. You can see I have lots to say the key thing is key differences are that industry is hierarchical organizations do have org charts right I have a manager whom I report to my manager has an executive that they report to. You know, the executive has to is accountable to the executive team is accountable to the board of directors of our organization. Being hierarchical in this way comes with being self aware. As I said earlier management is recognized as a entirely separate career track by transition to management. It's not I'm moving from, I'm effectively, you know, a senior I'm a, I'm a team lead a tech lead. I would be moving to a an entry level position as a first year as a first time manager. So, being hierarchical and self aware about it means that organizations, you know, there's can can formalize how they engage with power and power dynamics. I would add that the size of the market means that if an organization goes bad, if it goes sour, if there's abuse and to be frank there is abuse. You know, there's, it's the industry is full of people, and people will abuse power. The fact that there are many other opportunities. It means that people can means that people can go away from an organization and if an organization as a whole goes sour. Often people will leave, and it will shrivel and die. So that that's, that's very different from academia where and if an organ a department's culture goes bad. It is a much more difficult process to leave. And I think that's, that's a huge check on on power dynamics and and abuse. And I'll get to your question, why T and in a moment here, but I want to also point out that power and it's not directly addressed as power, but leadership and management training. All of these buzzwords are about the appropriate and useful wielding of power. You know, you can get an entire degree in it. That's what an MBA is about. There's books written about this. I do want to acknowledge that what I'm speaking about is my experience is white color labor where individual contributors are often hired into managerial positions. People leave managerial positions to go back to being individual contributors it's very fluid, there's no formalized dynamic, it would be very different. The distinction between, you know, an Amazon warehouse worker and a computer coder, a software engineer, very, very different dynamics at play there. So this is speaking specifically to power dynamics in white color in the tech industry. Why to your question which sector you think is more meritocratic, which more political, that is relations alliances matter for promotion assignments, etc. academia or industry. It's a good question. Frankly, I think industry is more matter meritocratic. In academia, you have to be. Oh, it's it's hard to say. Yeah, I would say that industry is more meritocratic. I think in academia merit does play a role, you know, if you're up for promotion in a research oriented job you have to have produced good work, great, or at least gotten published. And I would say that in industry there is absolute there are absolutely politics that is, you know, working relationships. You know, formal informal alliances, although I think that matters more at larger organizations. And for the four individuals in that managerial role who have kind of specialized in that work, because managing relationships is work, and it has to, it has to occur. And I would say that organizations in industry that don't produce value for their customers. They don't make money. They go out of business, right. So there is that Darwinian check it's much, much more removed in academia, where a department is really a small piece of a much larger corporation. In the industry, it's also easier to be shielded from from these politics by choice, or by by fact by focusing on doing good work and managing a relationship just one relationship with a manager, instead of needing as a manager to really manage a lot of relationships across the organization. I know that was a little bit of a rambling answer, but I hope it. I hope it helps a little bit and I should probably focus these thoughts and write up a medium essay or something. I would say that that academia also has. I don't think is very intentional in its approach to working relationships. And I think industry, at least, is much more intentional in its approach to working relationships and building those. So I have a lot more to say about power, but we have five minutes or so left so. I'll move on to the next section here, which is transferable skills. So that's selling yourself framing, framing your skills in a way that makes sense to people who are not academics, and then things to focus on. So, I'm going to talk very quickly about what you learned in grad school. Selling yourself that is being able to promote yourself to people in industry. And then ways to frame it frame your skills that make sense to what somebody in industry would be looking for. So, selling yourself. Basically, don't be shy. Make a lot of friends. You know, what one particular piece of good luck that happened to me was I became friends with a lot of the people in my department who ended up getting master's degrees and going back right into industry. Some of them who had come out of it. And that that ultimately if I trace back, you know, the network of professional relationships that led to my becoming aware of all state. I did a science job opportunity. That kind of it traces back to that so become friends with the master's students in your department, you know, figure out where they came from where they're going to go. And one other piece of selling yourself hiring managers, if someone is looking to hire, they're looking for how are you going to add value to the organization so it's the same skill that you flex when you write a paper. You care what I'm doing. Why do you care what I'm doing you know what who cares about whether you can hear the shape of a drum flex that skill when you're thinking about making a job. Skills that you develop as you get your PhD, things that are recognized, you've mastered your field, you've mastered your subfield you can take a research question break it down solve it write a paper about it, publish that paper. You know, who did what in the history of your fields you know the other people in your subfield, you know, what, what journals to go to to find answers to your questions who to ask things that are not recognized that are fundamental skills, finding questions, which questions are interesting or valuable to the rest of your subfield. That's probably true being able to evaluate a proposition as this might be a theorem this might be true or that's probably not and here's where I'd look to find a counter example for that. What's easy to prove, you know, knowing from the experience of trying and failing to prove something that oh that's an intractable problem right versus, you know, I can take this one. I can take this problem and probably take me a week to solve this. That's a critical piece of that's a critical skill. Here's some motivated project management, you can take a paper say I'm going to write this paper, break the this down into a series of homework problems essentially, and then solve each of those homework problems and wrap it all back up and present it back as a research paper. Lastly, communication, here's what we did. Here's how we did it. Here's the research program. Here's where this fits in that research program. I proved these two theorems supported by these limits. Here's how that paper breaks down. And this I think is is deeply underappreciated is that PhDs learn, they know how to learn, you know what it's like to be confused, and to push through that confusion, knowing that at the end of it, you will have more understanding. And that's a that's a critical thing that I think is deeply underappreciated about a PhD so knowing what it's like to be an expert, and what it's like to become an expert in something. Some, some practice here, research career tracks. See, talk to people who are in those career tracks, be aware when you're exercising some of those unrecognized skills, and then learn the new things that you need to learn for that career track. So it's, it's a career change. But being able to frame yourself as having a foundation that you can build on is critical. So, you know, for example, don't list your papers. Instead say you're a peer reviewed scientific communicator, right, you know how to communicate that's the skill, the skill is not the papers. The papers are supporting supporting evidence, you know, instead of saying I proved a theorem about this, say that you broke the theorem down into its pieces. Each of these is going to take this long, and then you push it all the way through. Instead of saying you know my specialization is global harmonic analysis. It's, I have a proven ability to quickly learn deep technical concepts, and I know how to use the research literature. I know, and I'm taking those skills and putting them to the test right today, even instead of teaching right, you're a proven technical communicator. It's a very to a non technical audience and you know that's one of the strong skills that I, I brought non technical skills coming out of out of graduate school at IU where I taught every year was the ability to teach to take something technical and break it down for an audience who's not technical. So, kind of think about framing from the perspective of someone who's not impressed with academic achievement but is impressed by the skills that you've built, especially those non recognized skills. Well, it's three o'clock, we've talked about everything. This is, this is the the end of my talk so here's a fin, or two fins rather. There's a little outline of what we talked about. I will put this in the chat again. Specifically the the document here that of with the spreadsheet this should be a link where you can go in and view it. In fact, the ams supports career transitions like this and it's trying to build out kind of a pipeline from from academia to industry and to support people who are making that transition. So that culturally it's not, it's a struggle in mathematics, but here are some resources that you can use also feel free to reach out to me as well. And I'll, I can be a resource as well, or try to help you find other people who would be interested in talking, if you have more questions about this. I have 20 minutes or so that I can stick around and answer questions if the if the zoom will be up. So, thank you for it's not a Ted talk but thanks for coming to my talk and I hope that you found it informative. Yeah, let's let's give you a round of applause. Wonderful talk thanks so much. Appreciate it. applause is so depressing in a zoom talk. Yeah. We have our all hands company meetings on on Fridays and, you know, we have what we have kind of an internal recognition for someone who is extra helpful that week and every week they get it and then you know it's it's exactly like, like you do like two or three people will unmute and applaud like you did Kim which I appreciate. But it's, it's kind of a running joke in the company now. I see a question in the chat here from from you Kim. So there is a different language. This is not the slide. Where was it. It's a it's a different language to a different zeitgeist for thinking about it. The framing here is so, and I apologize for for slipping into the the jargon. The critical thing is that when you put an achievement on an academic CV or resume. Right. It's a concrete output, a result, which signals that you have skills. In industry, you know, for the further a job is from basic R&D, the less they care about the specific academic achievements. And so, I mean, it's true, a PhD, you know, it's a great credential. It looks awesome. You know, it does open doors by itself, just by virtue of having it the status of it. But, you know, people don't really care that I know a lot about, you know, functional analysis, right. So, instead of the achievement, which means something to an academic focus on the skills that led to that achievement, which communicate potential to a non academic. So that's what I'm trying to get out with this is that instead of the achievement. Communicate the skills that you use to get to that achievement because those skills are what's transferable for an academia to industry. Does that make more sense. Yeah, I think what keeps catching me up is that these phrases that you use to me as well as an academic they sound like there is no content like there's no piece actually in there. That shows that the person saying this in any way has what they're saying it sounds like marketing speak. That makes sense. The, so here's one way, potentially around that. So, first of all, it is marketing speech right you're marketing yourself, but And most companies, at least in the tech industry will have, you know, be able to read through the marketing speech and say on the resume does this look like someone that I want to have a further conversation with. And then there would be a set of conversations where they dive into that like okay tell me about what it's like to be a peer reviewed scientific communicator. Right. What does that mean. And then you go into here's how this leads to the projects. One thing one way around that is, is to add the add the public at the output as kind of proof of it, but not assume that someone not familiar with academia knows what that output implies. So that, you know, like, you are a proven proven product product project manager, right, you can manage a technical project. Right. What that means is you can take a complex concept that doesn't exist. You can break down into a set of tasks, and you can do each of those tasks and then you can put it back together into a final product right. That's what that's what you do when you do a research project. Right. When, you know, you're proving a theorem, you have to have to work through all the steps and then back into backup for publication right. I'm sorry but the part where that doesn't exist is both key and absolutely hilarious. That's perfect. It's, it's, it's true right and it's the same things that you know if I'm building a new application, say to manage data transfers from one warehouse to another. Right. It doesn't exist I have to be able to envision it, break down into steps design it, put it all back together. So, and, you know, maybe platonically it does exist. I don't know, but it doesn't, it hasn't been written down yet. It's not known to be to be known yet so that having proved a theorem means that, you know, a big theorem or publish something means that you know how to do all that. And as somebody coming out of, you know, as somebody who has done the same thing, you know, I'm not a strong mathematician and I don't think but I have done math. I know what it means that you can prove a theorem. Right. But my hiring manager, you know, my manager who is a software architect, and has been a software architect for 25 years, doesn't know what proving a theorem means. So, it's kind of, it's there's a piece of education here about what are the skills that you pick up maybe informally to be able to deliver this academic output that you can take those skills, pick them up and drop them in a business. And the same set of skills will start going, and you'll be able to, to run with it and create value for the business. Jaffee asks, is there any difference between getting a PhD degree in data science and getting a PhD degree in harmonic analysis in industry. Good question. I think a PhD in data science would open doors, open more doors more quickly at, say, big research labs. Like, if you wanted to work at Google labs or DeepMind, for example, that that would open more doors, right, because they're actively engaged in that research conversation. You know, the PhD in data science means that you probably you've over the course of that conference, you know, you've been to NIPS for three years in a row you've talked to. You've talked to the people at DeepMind they probably know your face they've read some of your papers. So it opened doors in that perspective, as far as, as far as a kind of data science engineering job in the trenches, you know, the kind that I do or that I'm looking to, you know, that my teammates would be doing. My opinion is not. I would actually, I would treat them effectively the same because the set of skills that I need. Is skill, you know, it's, I need the project management and transferable skills, right. But the day to day skills are, you know, the are common to both degrees. So, you know, I don't most in a 99% frankly data science jobs, don't need a to know the details of neural net architecture, right. You need to know whether you've done this, you know, whether you've pushed forward the art of, you know, Monte Carlo identification of optimal hyper parameters. Most of the time it's it's a logistic regression, right, or a, even finding an interesting pattern in the, in the data and reporting it back out to the team so kind of those skills the ability to get data clean data. And that's, I think, common to all of those PhDs. Did that, did that answer your question. I know I ran my mouth there for a moment. I have a question that's maybe slightly tangential. I think at one point you were kind of talking a little bit about others, they're sort of more introspection industry about, you know, sort of the culture within. I think I agree where this is this is sort of the seat from a lot of, you know, academic departments, but I wonder what, like, how, how do companies do this sort of cultural introspection. Or, you know, like, what, what are they doing that's about academic departments. Good question. And to be clear, right. I just have five years of experience here. It's at two different companies. So I'm, again, scale, right, there's a huge variety of companies, some will be more authoritarian in the kind of hierarchy strict hierarchy sense. Some are very, you know, aggressively flat where there's maybe one leadership team and then 30 people who all report up to them. Or 150 people even where they're where. So, a huge variety of experience. I think what prompts these conversations is the existence of management as a career track, where there's people who are expected to their job is to make sure the organization is structured right. And the people in the organization are in the right frame of mind, having the right incentives to produce and keep the organization alive, keep it running. I think the fact that management from kind of the frontline manager that that junior first, first manager position up to an executive level exists as a separate discipline is part of why it has as a discipline has to create a language to talk about achieving its goals. And I think that the existence of that language is what allows organizations to be more or less introspective about the cultures that they're developing the hierarchies that they have the and and what's what's healthy and what's not. So, that's, that's my slightly considered opinion. And I think, I think academics have kind of the opposite conception is that everybody's departments very flat right, we elect the chair out of, you know, it's or it's your turn to be chair. You know, there's, there's no ostensibly no hierarchy other than, you know, from associate to assistant to full. Right. And I think that that is can be. There's a lot of ways in which that's positive right to not have an acknowledged hierarchy is creates a lot of opportunities for people. But I think at the same time. Any organization will develop a hierarchy. It's part of human nature. And if it's not acknowledged, it will, it will still be there. It will just be silent and enforced without without being known to be enforced and I think that's a trap that my opinion again is that's a trap that a lot of academic departments can fall into. So, that's, that's again, and to be clear, I'm not a management theorist, and I'm not a management expert. This is just kind of thoughts that I've had of, you know, watching organizations grow and develop and thinking kind of trying to ponder. So, it's not fully informed. And I am not, I don't want it to come across as as certain, you know, I'm, if you draw a management theory practice done in Kruger curve, I'm like, not in the, not in the asymptotic piece, probably better suited for sipping drinks, and at, at a happy hour, then, and, you know, answering authoritative questions at a, at a lecture, but take it for what it's worth and take it with salt. All right, so if you have any other questions, will we wrap it up for the day. So if not, let's maybe thank you all once more. Very much.