 First of all, sorry, I've been talking a lot, but I haven't introduced myself. I'm Ewa Takouk, co-founder, CEO of Feedback Foods. Also in the call is Irmak, she is our head of marketing that arranged everything around making sure that I would be able to pitch here. What I'm going to tell you about today is something that it's a it's a R&D project of Feedback Foods. So it's not a product that Feedback Foods has or is selling. It's an R&D effort that we did in collaboration with the TU Delft and with MITx. But before I dive into that specific project, I just want to tell a bit about Feedback Foods. So Feedback Foods is a spin-off of the Technical Diversity in Delft. Our core mission is to improve the learning designs and learning experiences institution for higher education. Because we believe higher education is this pedagogical transition and where for the first time the monopoly that universities used to have is being challenged because companies are starting to hire based on skills more and more compared to hiring on just only diplomas and there are other companies starting to compete with universities as well. Why do we care about this? We care so much about this because the university is so much more than just creating soldiers for the workforce and creating these skills to get people into the workforce. It's a fundamental element in our design of democracy in our social design. It's the only institute that also prepares future leaders with the mental tools and the morality that is needed to drive basically our societies forward in the right direction. We care very much about that. This makes us a bit of a different type of company where 80 people now operating in US Scandinavia mainly in Europe and of course in the Netherlands and also in Australia. However we always wanted to maintain our independence so we're bootstrapped because we really want to focus ourselves on solving this problem. Hence the fact that we have an R&D team working on technologies that I'm now also going to share with you is because we feel that there is so much that we still need to discover that plus we shouldn't only build a tool that sells we should build a search engine that is continuously looking at how we can improve these learning designs such that will lead to those cultivating of these skills that we need for our democracies. Well this is maybe a bit of a fake story but I think everyone I hope it resonates with the people whenever it's talked about like that. So Peabock Foods we build tools to support engaging learning designs but we are a mission-driven company and we have an R&D team and what I'm now going to show you is one of those mission-driven elements because and I don't think I'm preaching to the choir here but the role that publishers have in education is definitely something which I think is part of the past should be part of the past. There's a lot of good content out there very high quality however there's still this gap between this high quality content and a student that is now studying for an exam and needs to pass it and has just another couple of days to study. It's this sort of sweet spot that we tried to solve for so what I'm going to do is I'm going to share my screen and well I'm doing that I want to give you a bit of context so this is a project that we've worked on two years ago as I said before with MITx and TU Delft so unfortunately because we weren't able and this is the full-provoking element that I would like to discuss with everyone in this call later because we weren't able to find a product market fit for this product we sort of have put it on the shelf so we do not continue we are not investing in development anymore which I think is a big shame because it can be of value so is it correct that you see a YouTube video here sorry that I have to use a YouTube video here but I said it's on the shelves so we don't have it so okay so this technology what it does the core idea is it takes the 100 accredited sources of open online open course from taking MITx as one of them Harvard all those ones that we know the material is at least curated and good enough to be shared and then the type of material it's videos explanatory videos but it's also exercises and well everything so the use case here is a student that has a learning linear algebra exam that uploads this exam to this system and then the system automatically extracts the topics out of this exam so here you see okay we automatically see it's about linear algebra factor space system of linear equations abstract algebra etc and then it searches through these sources in order to suggest open courseware that is explaining these specific elements so for example now you get this lecture six of MIT in linear algebra spring 2005 so it and our next the next version that this video doesn't capture also shows you the video and then exactly in the video tells you okay this is when it is talking about for example linear algebra so if I think I have that yeah so it jumps right away at the moment of this video and when it starts talking about specifically for example one of the core features that sorry the the the topics that it has so what it does how we how we the technology works here is as follows so we what we do I'm making sure that I yeah so we have we built this engine that first of all our break breakthrough has been the ability to build build a knowledge graph so understanding not only how our topics connected but what is the hierarchy between those topics and that's what we refer to as a knowledge graph then what we do is we take these repositories of material and we map them on this knowledge graph so where is it and then what we do is we take that examine this case extract those topics and then map it as well on that same knowledge graph and there where we have the most overlap that is the that gets the highest search results so what you see here is a design that allows for different types of use cases so the use case that I just showed you is an actual standalone website and where a student uploads it but what we've also built is an integration into existing management systems well we copy that from our sort of normal tool suite because we have this integration available but then it goes into this course it analyzes the information that's already there in the course so for example how are weeks structured what what are the types of the weeks is there already material out in those weeks then we scan that material in order then to suggest online open courseware for that specific course where the use case is a teacher wants to create more learning paths not just one because well this is a trend right we want to have more personalized learning but how is this teacher going to find all that material well automated through this through through this system as well as a browser plug-in and where you could you know you're you're here you're in a plane geometry geometry in Wikipedia and you have this plug-in saying hey explain me what is here because I can read it but I still don't really understand and then it suggests this open course rather actually explain the looks that are described here so this has been a project that we've as said worked on quite intensively with quite a lot of people however we've never been able to really build something that sort of this this way has been really a technological endeavor us being more an engineering company coming from Dewey Delft working with MIT hence the big pitfall that that we could make is you start from the technology perspective that's what we did so I think we've made quite some technological breakthroughs but we've never been able to really find that product that could that actually took off so yeah I guess that's where I am and I guess the open part of this session is I would definitely want to hear your thoughts on on this and also on the fact that it's now actually laying on a shelf getting dusty instead of actually being used for where it has been designed for so yeah well back to the studio I would say thank you yeah wow I hope this is thank you yeah yeah no I think it's very interesting the way as I put in chat the way that both searches a curated collection in this case the MITx material and then enables you to quickly jump because of course the thing about all the repositories is that it takes a huge amount of time to sift through all that stuff to find something that's relevant and useful for you and it feels like this is a tool that's helping you quickly access curated materials and then find the relevant things that pertain to what you're trying to learn I guess it's it was a bit of a jump for me though to go from the MITx material to wikipedia so I got you sorry so can you explain how that works like because I think that's also very interesting in terms of drawing on additional content right beyond just the curated collection of MITx with your poll so I think and I did not explain this very well so we're able to tap into all types of resources right but what I wanted to highlight here is the big breakthrough has been this this technology so this knowledge graph and the way that we able to map it but then in the front end so the use cases that we support they can be very variety of use cases one of those use cases is that browser plugin that I was mentioning where it does exactly the same thing instead of analyzing a document it analyzes this web page in order to suggest your material to explain it in this specific scenario and here it analyzes the document that you drop here in order to explain it and this other scenario is a plugin into the learning management system and then takes the information from that course in the learning management system to do the same trick so this is actually what I wanted to show there's a lot of different types of use cases that we could think of however for us it's hard to understand we didn't we did not succeed in actually getting to that use case which took off in the market and therefore we did sit and we continued with another AI project which is now taking off like crazy so yeah also interested in seeing if we could help others out with this technology can can I ask um so what's been the stumbling block uh from sort of I might have missed it but um uh from putting putting this out there at the moment like why are you sitting on a shelf gathering I got I got you well there's a couple of reasons to this uh our sort of go-to-market strategy was replay so if you go to MITx open course website as I think 100 000 unique users every month our core our go-to-market strategy was replacing that search engine with this one however I think I have to be polite here in saying it the right way but I hope no one feels offended there was more to it than only technology to get that uh landed and unfortunately it didn't therefore so we sort of lost that that battle then we tried another way which was actually having it as a as a a browser of a learning management system plugin but then we just saw that teachers didn't so we have our logo is sweet so I'm sorry a fruit basket and we have a lot of different types of fruits and we'll check which fruits do people pick this one wasn't just a lot by teachers and then I guess it's sort of it drained our basically it is r&d it was in an r&d phase so because we couldn't find that product market fit we jumped on to the next thing and hence this is on the table however we still love it there are engineers in our team that still sort of have a big heart for this but yes we're just um our distraction goes too much away to that new project that actually is taking off can I can I ask if um because I might have an idea for a funding solution for you um is the software that you're building it on is are you happy to have that is there the reason why is because grant for the web is giving out some big grants that you could kind of put their monetization tool on this so that if people wish to you know monetize like you know to to sort of give back to the project they could um and I'm I'm wondering if um I don't know if it would fit or not but but there's been some pretty huge grants they've given out to some companies that are that are creating you know open tech yeah I'm with you yes we're open for this for sure it's it's it this is not something that our core business model is relied on relying on so it's and it's it's a way as if it's on the shelf and there's a lot of effort going into has been going into this to see if in some sort of extent we can make this ready for to be open sourced and then open source it yes there are already some sub libraries of this project already open source if you go to our GitHub page of feedback foods um but please let yeah let's let's let's get in contact I'll make sure I'll connect you to our lead engineer um and we'll take it from there I should specify I'm I'm not with grant for the web but they've given me a grant recently so I can I can pass you along all the information on how to get it to get in touch with them that that that I already understood yes okay good no no don't worry don't don't worry about it yes yes yes I'm I'm not sure exactly Erica where but what your background is um I'm a storyteller and an educator and kind of a self-taught um tech geek uh I I don't do the level of things that you do but uh but but I I I do sort of dabble and and build stuff