 Hi. My name is Mike, and I'd like to tell you about the future of machine learning. So a Deloitte study shows that the number of machine learning projects piloted in the world is going to quadruple within the next two years. I'm sure that's not news for you guys in the crowd, but the fact of the matter is that human brains need to build all these algorithms, at least for the time being. And there's a giant skill gap between the number of machine learning engineers that exist in the world and the number of projects that are going to be demanded. So the number one source of machine learning developers in the world is actually software developers that are re-skilling on the job. The problem is that for these software developers, machine learning is still a black box, totally different animal compared to software development. We did a large survey before going to market with over 100 AI companies, and we found that over 80% of them, for example, use Slack to make important decisions about machine learning. And yet, those decisions, when they're made on Slack, are really tough to carry out. What does that mean? We got a billion dollar problem. Why? Because when you make the decision to turn a server off on Slack, when you make the decision to redistribute compute power on Slack or kill off a model, there is no actionable way to take that decision into your testing environment, let alone in production. Our survey estimates that this leads to 10 hours of waste per engineer in a company in machine learning stack and development process. So at Mishkin, we decided to fix this problem with the world's first interactive machine learning bot. It is IML software for the tweak, test, and learn generation. Before I tell you about how it works, the entire team would like to thank our clients and our partners for helping us iterate this tool in beta before we release it to the world. We worked backwards from client need in machine learning to come up with an amazing, uncharted list of engineering problems to solve, and solving them through our product has been a wonderful adventure. So let me tell you about what it does. First, with our interactive machine learning product, out of the box and for free, you get monitoring of all your machine learning jobs on Slack. On top of that, you're able to schedule as many jobs as you want right from Slack, pause longstanding jobs that are executing, tweak parameters while the job is executing, fork a job just like you'd fork a repo on GitHub, and under the hood, it will automatically spin up, provision a server, and send the job off running. Looking at our competitor matrix, we are the only company currently that has a duly UX centric and a deep tech approach. We have a patent pending for the optimizations, scheduling, and memory IO options that happen in the back end, which are purely a result of the requests that we get from our front end application in Slack of what machine learning developers want. On this subject of what machine learning developers want, our business model derives entirely from that. We have the privilege of partnering with Slack, Microsoft, and Amazon for their machine learning workshops, which are conducted on Slack through the Meshkan tool. That means that we're there at the workshop feeling the pain point of hundreds of machine learning engineers as they're discovering these tools for the first time, interacting with them on Slack, and turbocharging our sales process directly in front of the clients that matter most to us. Of course, Slack is an ambiguous tool that allows us to give high-level support. And furthermore, most of the advanced features that you saw us talking about are part of a freemium model that you have to pay for. But in order to measure what value we're capturing, we use analytics tools directly through our partnerships with these providers in order to see the value we're capturing, charge for it, and continue to delight our customers. So our market is barely out of the gate, and already it's a huge business opportunity. We seek to capture 20% of the total addressable market by the year 2021. And as I told you guys, that doesn't just stop with machine learning. Because remember, these are reskilled developers who also need to test their code, push their code to production, and mostly be advocates for a data-driven approach in the future. We hope to be their trusted source of software tools now so that we can scale with them as they grow. The Meshkan team is myself and three intrepid entrepreneurs that combined have experience of 16 years providing deep learning and high-tech products. We build our solution with an immense amount of love and empathy for our clients, all of whom are colleagues, research partners, friends, or future friends. They are serving all of you in this audience with new generation machine learning tools. And we are proud to serve them. Lastly, I would like to thank our investors, including Risto Silasma, chairman of Nokia, who are supporting us in our journey to be the world's leading provider of machine learning and development services for a new generation. Thank you. Yes. Thank you, Mike. Jury. So very cool, great presentation. My question is, the machine learning number of engineers, they were small, right? And we heard that from Amazon just one presentation ago. How big of a business can you build here? So this question leads me to an exciting announcement that we just launched a second product based on our machine learning stack called Unmock.io. Don't take out your phones and type in Unmock.io that uses the same exact stack for continuous integration. So let me rewind a little bit and talk about what I mean by tweak, test, and learn. When you're running a machine learning model, it's a very data-driven process. So you're tweaking it based on external data, but not based on lines of code you're writing. The same exact logic works for third-party integrations, where you're tweaking code and testing it based on mocks of third-party APIs. So we found that machine learning is the first vertical that we're scaling up with, because we know it well. I myself am a data scientist. And by validating this tweak, test, and learn model, we hope to catch a much larger total addressable market, including continuous integration, monitoring a production environment, and database management. So are you providing data then for that tweaking? Yeah, so actually, the second service I talked about does the first service with machine learning, the developers bring in their own data to the servers. And then the tweaking, testing, and learning is actually about the configuration of the models and the parameters that go into those models. I have one comment and two questions. First, you have a very special company name and logo. That's very special. Two questions. So how does onboarding, if a customer looks like, is there a customization? Or it's like an automatic or straightforward process? Yeah, so that's another question that could be answered with a website, so app.meshkan.com. And we could see how many of you are onboarded pretty soon. So it's through Slack. So you go there, and immediately you go on the Slack page, where you can integrate it into your favorite workspace. You'll get a ping from the Meshkan bot, and you could take it from there. The other question is whether you're optimized for certain workloads, certain machine learning specific workloads, or it's like totally generic. So we're most optimized for deep learning workflows. And to tell you the truth, it doesn't make a lot of sense to use a tool like this for short running workflows that don't need a lot of monitoring and constant re-tweaking. But for deep learning workflows written in PyTorch, Keras, SCI Learn as well, and TensorFlow were optimized. And we are big users of Meshkan as well. And you can see that in our open source repositories in GitHub, where we'll show you how to use the thing. And how are you setting up your sales team? So our sales team is currently the entire team. I was just at Reinvent last week, where I was there both to pilot these workshops and to sell. But I read a great book earlier this year called Negotiating with Giants. I don't know how many of you guys read it. And one of the great pieces advice that I learned from it is sometimes putting on the other team's jersey helps a lot to be able to sell stuff. So I'm proud to say that our partners are part of our sales strategy, and really actively helping us with that. So Microsoft for startups, AWS for startups, it's not just free credits, but it's a whole lot of advice about how to sell. So it's a win-win both for the providers of the GPUs and the people like us that are helping you use it. Continuing on the sales, you're targeting the companies and teams in the companies that you're not selling to individuals. So currently, we have a dual strategy. We started by selling to companies. And those we meet at conferences. And I've met several of your conferences, including this one. But after beta testing the tool, we've moved to selling to individuals. And then upselling through the Slack bot. So we have instant onboarding. And then the upselling is all behind buttons that people click on, which then allows us to measure price and adjust price on the fly so that we can make sure that we're always providing value to our clients. So I gather you're raising financing? Our financing. So currently, are you raising capital? So now, now we're not raising capital. We are going to hit the road in San Francisco in a month to start doing that. And if any of you are investors out there interested in financing Michigan, we're especially looking for people that could help us scale with this model. So we'll be talking to the Slack fund and a couple other funds, too, that you can imagine that the Slack fund knows something about Slack. Yeah, I would gather. I've also, we've seen platform dependency be a killer of businesses in a sense. What kind of a dialogue do you have with Slack right now, right? With several sources. There's CTO who might even be in the audience is there, is here. So I had a chat with him last night. I had brief chats with the sales team as well. And I'll be there in San Francisco to meet them next month. And I'll also be at Amazon headquarters in Seattle. So the dialogue is pretty fluid. I think everybody realizes that it's a win, win, win for all businesses involved. And I'm really excited that they're willing to help us out. And I'm also kind of excited to help them out, even though we're still small fish. OK, I can ask more. Yeah, one more. Where do you aim to be in 18 months time? In 18 months time, we would like it to be deployed on a thousand different teams platforms, running machine learning on Slack, have data-driven metrics to figure out the pricing model, and have launched two other products based on this tweak, test, and learn model, going with this chat ops idea. So you already saw the Unmocked product. And currently, we have one in the works for database automation as well. That's where we would like to be in 18 months. Those are some pretty good milestones. What does that mean in terms of are you going for a freemium model that you're going to deploy? So everything currently runs on a freemium model. And we also meet with larger teams to come up with custom solutions for those teams. For example, I'm flying out to Stockholm to meet up with Raysearch Labs next week on this exact topic. So if you're a larger data science team, don't hesitate to reach out, have a chat. We'd love to come up with a custom solution for you as well. But in order to be able to scale, the nice thing about Slack is it's already in millions of pockets, thousands of which are vibrating around this room. And we hope to sell through that too. Can I just give you a compliment? I think you're an amazing marketer. Can we just give him a round of applause for that? Because you can. Thank you, Mike. Thank you. Thank you very much. Take care.