Max Jacobs of DeepMirror says AI could cut drug development from four years to six months. The platform aggregates algorithms and keeps customer data isolated in a certified cloud. FDA approval still requires animal studies, but the AI work beforehand faces no regulatory barriers.
Chapters
It takes a really long time to translate drugs from the lab into the clinic, an average of around 4 years or so. And we at DeepMirror bring that down to 6 months with our interactive co-ideation platform, which essentially is a tool that helps chemists and researchers in general to collaborate with AI and their colleagues to design safe and effective drugs faster by helping them suggest the next best molecule to make.
I'm Max Jacobs. I'm the CEO and co-founder of DeepMirror. We help chemists get drug molecules from an idea into patients much, much faster than ever before. In academics, like, they often work mainly for themselves, for their own prestige. So it is actually quite isolating and quite lonely. So you don't actually work often as a team. The last one is the professor, the first one who did all the work. Everyone in between doesn't really get much of the recognition. So teamwork is actively discouraged in a way, whereas now in a company it feels like everybody has a specific responsibility and works as a team. It feels like you have some sort of perpetuum mobile that kind of moves by itself, and everybody does their stuff for the greater good, which is very different to academia. I mean, I'm still a scientist by heart, even though I left academia. Initially you think, oh, you just do it for the chemist, da da da, but then as soon as you feel it touching other people, and see what it would mean in the grand scheme of things, then you know actually the really addictive thing is the patient impact, not so much the steps in between. But you have to kind of do it the whole way. Being on that track, you don't feel like you're absolutely in control of your own life most of the time, because in the end, scientific research can turn out really well or really bad, and you have very little control over that. That's why I actually kind of shifted as well. So it was a bit of a gradual thing, but it came from the idea initially that it was all about trying to translate technology into something that people actually use every day in their work and not something that they don't really understand how to use or it's too complex to use.
CEO, co-founder at DeepMirror. Each and every one of us here in this room will have 20,000 ways in which you can catch cancer that all need to be treated by different treatments. However, over the last humanity has only approved around 1,000 drug molecules for treatment, and we need to do much more. We just sell directly to the biopharma companies, and we tend to aggregate algorithms that we know work well. We keep also doing our own research on top, but first we test.
John, do you have questions for him? Is the FDA fully on board, or does it not even touch them because, you know, you're, you're all kind of pre-FDA work?
I don't think we have any regulatory issues because in the end, for FDA approval, You would still need to do all the relevant rat, mouse, and other animal experiments. So that's, of course, super important.
You're paid just for the service, or do you get a piece of the drugs that get designed on your platform?
It's like the crux of like where we are different. Many investors either go on one side or on the other side.
As soon as we would take a piece of the drug, then it becomes a bit competitive because then suddenly we take a piece of one drug and then suddenly we become a biotech company ourselves. And then the whole like data sharing part doesn't really work that well anymore. 'Cause people might not trust us as a data arbiter. So it's a subscription service primarily.
As you get these customers and they are feeding in their research into your systems, are you isolating those data?
It's a multi-tenant cloud, but of course there can be no cross-contamination. So we're actually ISO certified and also in the process of getting SOC 2. I mean, that is super serious, right? None of these molecules can ever go into the public domain because each of them could become a billion-dollar drug. So there's a lot of, of a trust element involved, and we're taking a lot of effort. To protect the IP.
Max, thanks so much.
It did seem like there was a moat here that I got, especially based on the data that they're getting from their customers. I think it's a great space. Like, who doesn't want drug delivery to be delivered faster and to cure malaria?
What'd you think? We didn't get enough into what the subscription model looks like. I thought he made a good comment about wanting to make sure that they don't favor one company over another or they don't have a business. But, you know, I don't know how big a business this can be. If you're not taking a piece of the drugs, I mean, that's what the real business is. How do you beat that?
One of the obvious things to do at some point would be to hook up a lab behind the platform and then over time become really like the main platform which everyone in the industry develops drugs. And frankly, nobody has ever done that before. Maybe the closest A good comparison would be something like BenchLink, which is a company that is now $6 billion.
