Office Hours: I am an AI resident at Google Brain and venture scout at Backed VC.Featured

Hi everyone! I’m an AI resident at Google Brain in Zurich, conducting research in transfer learning. I am also a venture scout at Backed VC, a founders-first seed-stage fund based in Europe. I blog about machine learning (ML) and how to learn ML at I had a fairly non-linear path to where I am now - I started out studying Maths at Cambridge with the intention of becoming a pure mathematician, but then took a gap 9-months after realizing I wasn’t so sure and wanted to see what working in the ‘real world’ was like. In that time, I interned in digital marketing and eventually taught coding, founded an EdTech startup and wrote a book about startups in Hong Kong. As I resumed my studies, this time in Economics, I became interested in machine learning and started learning through Udacity’s Nanodegrees, eventually doing a quant research internship at Jump Trading, switching to study Engineering and then doing a research masters in machine learning at UCL.I enjoy learning languages and playing musical instruments from the cajon drum to the bass guitar. :)Ask me about learning or transitioning to AI/machine learning, doing research, early stage startups and investing, learning languages or something else!Update: Thanks for the questions everyone! I will continue replying to them tomorrow. Feel free to ask more questions that you have on your mind - I can't guarantee I'll get to them, but it never hurts to ask. :)
Thanks so much for joining us, @jessicayung. Elphas: please reply in the comments with your questions for @jessicayung before this Friday. She may not have time to answer every single question, so please emoji upvote the ones that interest you most. Thanks!!
Thank you for your inspiration to others! I love being on startup panels and was just on one for executive mba students from Austria. I would love to hear more about your European seed fund. Thank you!
Very excited about this office hours! I haven’t been on Elpha in a while (great to be back, and congratulations on the app!😊🙌) - can you remind me how it works? Is there a livestream element or will she just be answering a few questions directly in this thread? Thanks a lot!
hi Jessicawhen you say early stage, what exactly do you mean? I have an early stage, deep-tech health startup. We are collecting data and developing our AI models, anytime I talk to funders they say we are too early stage. We have some customer validation but not a lot as with healthcare we need to have the solution developed before customers properly engage with us. It's really tough to navigate this line, would appreciate advicethanksShona
I'm facing a similar challenge @shonadarcy, @nancyahola, @jessicayung: we've built a complete DNA analysis platform for optimized medications management but tech investors worry about things they don't understand (e.g. genetics, drugs) and life sciences investors can't et comfortable with the data moat (instead of traditional patents). Plus a very small pool of investors in Canada, where we are based, limits our reach. Have customers, will travel!:)
@shonadarcy I think with the change in COVID - dominated atmosphere right now, healthcare is extremely different, even from a few months ago. Don't talk to funders that are in the tech space - speak directly to ones in the healthcare space. I work in that area and the funding conversations are happening far more than before. Feel free to reach out if you want to know more. Happy to give input.
Hi Shona, great question. And thanks Nancy for your thoughts.First, props to you for persisting through what seems like a very frustrating process. I've written some general thoughts below, but it'd be great to know more details about what stage your startup Kids Speech Labs is at and what investors you've been approaching to understand where you're coming from better. :) With that caveat, I'd say 'early stage but not too early to invest in' depends on the kind of startup (as well as the investor). For example, for a consumer product that requires very little technology, many investors would look for strong traction or customer validation. If a lot more R&D is needed to even make the product really 'work' or help solve the problem in the market, such as in healthcare or biotech, investors might be willing to invest before one has much - or any - customer validation in terms of actual clients. There has to be strong evidence that there's a market for it, of course. But since this can be riskier since (1) there isn't proof that the exact product has got traction yet, and (2) there are more ways for a company to 'fail' even before they go to market (though in this case going to market is a much bigger win), some investors may prefer not to invest in such companies, especially if they don't have expertise in that area.*Concretely*, I would say:(1) try to find investors that have a history of investing in companies in deep tech / healthcare or other areas that require more R&D, so they understand that it takes time to get to a product. I.e. similar to what Nancy said. Investors that understand your area will also be able to support you better later in the process, which is important. :)(2) Make sure you have strong evidence of customer interest or there being a market, since the additional R&D investment period may warrant stronger evidence here. (2b) An alternative could be trying to get more detailed feedback from a smaller trial group - very strong feedback from early adopters can be as good as general traction from a larger number of customers. E.g. you mentioned studies on the Kids Speech Labs website - perhaps you could see if more parents would be interested in trying your proof-of-concepts out? Of course easier said than done, but my guess is many parents would be interested in getting more information on how their children are developing, so it may be possible to get more trial users even if the product is not nearly complete. :) Another alternative is to get feedback on a low-tech version, which may even be your team manually doing what you hope your tech will be able to do.Another thing that may help is (3) strong evidence of technology developed (exclusive access to data counts as well). For example, one of the startups Backed funded had a strong technological proof-of-concept though it wasn't nearly complete. If I recall correctly, they had no customers, but the market potential, technology and team were super compelling, so Backed was keen to invest. :)Hope that helped and good luck!PS, I am also a big believer in the importance of early education, looking forward to seeing what you do. :)
Hi Jessica, thanks for the chat. Any advice for a 50+ year old woman with an AI company Hello Career Guru a career trainer to help women advance professionally-looking for funding? We plan to expand to Europe after the US. Is Europe a good place to look for funding especially since Germany and several other countries require corporations to have women on their boards. thanks!
Hi Suzanna, Hello Career Guru looks exciting. I would be interested in trying it out. :)Hm good question. I'm not too familiar with later stage investing - my guess is that the answer to your question depends on the specifics of one's company and stage.The fact that several countries require corporations to have women on their boards (cool!) definitely helps, but seems more like a bonus than something that, alone, is enough to propel strong investment.Why not start talking to potential investors based in Europe early and ask them what they think? Even if you're not hoping to raise now, it helps to develop relationships early, both for fundraising as well as for growing one's network.
Such great advice. Love to chat how do I reach out?
Message me on Elpha :)
Hi Jessica - I have two related questions I'd love to get your thoughts on. First, what do you think is the most interesting NEW area of AI, and second, what do you think is the most interesting OLD area of AI that everyone seems to have forgotten about?
Haha that's a good question! I'll interpret 'new' to be 'not widely adopted' because some would say most recent developments 'originated' at least 20 years ago. :)I am very excited about generating or simulating data because if we can do that, we can have an almost infinite amount of well-labeled data which we can then leverage. There has been much progress in this area recently, so we'll see what happens.It's hard to come up with an old area of AI that everyone seems to have forgotten about and that I'm keen on, in part because there are so many people in AI now that most things are being worked on in some capacity. At the moment there is a lot of combining of deep learning with other ('older') ideas, in particular to make previously unscalable algorithms (e.g. many Bayesian ones) scalable. It's all been cool to watch. One of the most interesting topics is causal inference: trying to make models learn to reason causally instead of simply learning associations. It'll be interesting to see how much or whether ideas from 'old' more-symbolic AI may be interpreted and adapted to mesh with deep learning to do this.
Hi Jessica! So nice to hear about your story, especially that I'm now based out of Switzerland too and was actually born in Hong Kong :-) I've mostly worked in the corporate world for the past 15+ years, but the entrepreneurial bug in me has always been itching inside. I've been planning for my jump into tech and VC. Would you have any resources, VC firms, any starting points to look at in Switzerland or other places in Europe? Any particular firms, industries, or positions you'd recommend me to look at with my sort of background to make such a move? Would trainings in engineering, AI, machine learning or the like a must to enter this world, and specifically in Europe?
Hi Venus, Oh wow cool! It's a small world :) What roles or parts of tech and VC are you hoping to go into? And what have you been working on in corporates?Training in ML / engineering is definitely not a must to do VC, though tech is a different story depending on what you're hoping to do there. For VC, I think it is definitely worth understanding at a high level what ML can do now, what it's likely to be able to do soon and what the limitations, risks and issues are. Having more knowledge in ML will likely make it easier both in the sense that (1) that gives you an edge and expertise that not many people in VC have, and (2) it will often be useful especially when evaluating and working with companies that use ML. So beyond a high-level understanding, whether it's worth it depends on how interested you are in the area, how easily you can pick it up and whether there are opportunities you think you'd likely be able to attain after you get that skill.In general on USPs, I think that a good way of approaching it is to build on your background and current strengths, and think about how you can adapt or build on them to offer a unique perspective. E.g. if you have more corporate experience, perhaps you could give more insight on B2B if that interests you. And if you'd like to combine it with AI, perhaps you'd have a better insight on how corporates might adopt technology and what the difficulties and opportunities are there.As general practice, I think that evaluating deals really helps develop 'VC skills'. You could even start a blog on that to practice, show your interest, build a portfolio and maybe even get some feedback. :)
Thanks so much for the directions Jessica! It's clear, I believe what I have to do next is really to get some overall good understanding in ML and AI to get started. My corporate experiences are mostly around digital businesses for 'traditional' or pre-existing brands - media, advertising, retail. It's been often 'catching up' or 'rectifying the past' and now I'm really longing to participate in building the future. VC is exciting and I'm passionate about supporting women entrepreneurs so I figured having influence over how checks are written and to whom they are written matter a lot. ML and AI are all important and critical to our collective future so we must be know enough to make sound judgement even when I don't aim to be an expert. Regarding starting an evaluation portfolio, would you have some publicly viewable (i.e. learnable) recommendations?Thanks a ton and I hope you are enjoying the great weather in CH these days.
Hi Jessica and what an interesting journey! How best would you recommend following up with VCs I pitched who seemed interested but haven’t given me a final yes/no after following up with addtl materials they asked for? How does one “nudge them along” politely?
Hehe personally I'd say just nudge them (as long as you don't say something explicitly rude of course), especially if they showed interest. They are probably just busy or forgot.Persistence and following up with important clients are hardly things that I would rate a team poorly for. :)
Hi Jessica!Would love to hear more about how you became a scout. I'm constantly meeting early stage founders and also know some folks in venture and would love to stitch these all together by becoming a scout. How did you break into investing and become a scout?
It's a funny story actually. I was first introduced to Nathan Benaich (an investor, now founder of Air Street Capital, that also writes a fab newsletter on AI x investing and does lots of great things in the space, look him up by a friend at Udacity. Nathan then introduced me to Pari, the founder of The Engineering Company, a startup he'd invested in, so I could introduce him to talent in Cambridge, UK (where I was studying at the time). Pari then introduced me to Backed VC (who'd also invested in Pari's company) when they were looking for their first group of scouts. So it was really a series of connections over time combined with luck and being a good fit at the moment the opportunity came. :)As with many things, scouts are often found initially based on recommendations of people close to the fund, e.g. existing scouts or founders. With Backed, a few people have also just reached out (e.g. to existing scouts) directly about wanting to join, and I think that can also be a good way to go about it.I think the best way to get feedback on your situation specifically would be to reach out to a fund that does have scouts and ask and get feedback. And well if they say yes then you're in! If you're based in Europe feel free to PM me. :)Two key criteria are (1) having strong networks or awareness especially in places the fund doesn't have much reach in, and (2) showing interest in the area. Of course (3) deal-evaluating skills are a plus, but I suspect that is less key than (1).Good luck!
Hi Jessica, thanks for doing this. What does a AI resident at Google Brain do? How do you apply and qualify for it? What do you get out of it?
Thanks for sharing your story! As a current undergrad rising senior, I'm wondering how you joined a research team at Google and connected with a team in Switzerland. Do these positions typically just require a masters (or a special research masters?) or is a PhD necessary? And did you meet a recruiter from your current team while at UCL? Thanks again!
Hi Jessica, Thanks so much for taking the time to answer questions on Elpha. My questions are about the residency program at Google. (1) Do you think you were prepared enough in DL/ML/SWE before starting the residency program? If you struggled in some areas, do you mind sharing?Do you think having conducted research in ML/DL prepared you for the rigor of the program or do you find people directly coming from industry can manage just as well in a research setting? Is there a striking difference between recent graduates versus people coming from industry who are deemed successful in the program? What skillsets and mental attitude do you think are most vital in performing well at a residency program?(2) Do you feel that you get enough support, guidance in your research/project in terms of professional growth and long-term goals? I assume everyone is quite busy and focused on their tasks but do you think supervisors are generally there to guide you at a distance so that you deliver a result or are they also invested in your career growth?Thanks!
What are the main struggles of a first-time startup founder?
Hi Jessica. I love that you're a girl drummer and bass player! I am an experienced Entrepreneur who's raised capital and had a prev successful startup. Currently, I'm a non-tech founder of a startup re-inventing shopping using AI and deep tech. Do you have any advice on where to find a technical co-founder with your type of background and possibly even get a referral? I found two guys online that had amaz'g credentials on paper; but when it came down to it, they misrepresented their skillset and I had to dissolve my corp documents and agreements and now am starting over. I wonder if there are more trusted sources for great people, etc. Thx so much!
Hi Jessica, Thanks so much for sharing your time and expertise here with us! As a COO at an early stage startup that is in the travel domain, we've been heavily impacted by the current situation (as you can imagine). Just two months ago we had investors knocking on our door, and I believe we still do have a very strong product in a terrific market that, if trends continue, is only going to explode. However with the uncertainty and inevitable changes that are about to happen, I'm wondering what our best approach to raising funds would be at the moment. The thing is, the tables turned - where we were the ones who could pick and choose the investors based on our criteria as cashflow was steady, we're in a situation now where we need to fight for survival, but I'd hate that to dictate the terms so much that we're forced to make compromises we're not comfortable with just to raise money to survive. Any ideas on how to tactically tackle these talks now? Thanks again!