BIOS Community
Published in

BIOS Community

Founder Spotlight #28: Max Jakobs @ DeepMirror

BIOS: Nucleus of Life Science Innovation 🚀





Alix Ventures: Supporting Early Stage Life Science Startups Engineering Biology to Drive Radical Advances in Human Health

Calling All Innovators Click to Reach Out 🚀

Max Jakobs is Co-Founder & CEO @ DeepMirror. A theoretical physicist by training, he went on to do a Wellcome Trust-funded Ph.D. in Neuroscience and Machine Learning at Cambridge University in the UK. During his Ph.D., he built software packages for academics across the globe and Co-Founded DeepMirror to promote the widespread adoption of production-ready AI in the biomedical sector. He holds a BSc and MSc in Physics from Cologne University and an MPhil and Ph.D. in Developmental mechanisms from Cambridge University. Max is passionate about making AI usable and relatable to revolutionize the way we do research and treat disease.

DeepMirror enables the creation of biomedical AI for the whole life science cycle from basic research over pre-formulation studies, to clinical trials and diagnostics with a single technology platform: DeepMirror Spark. DeepMirror Spark requires up to 100 times less data curation to grow AI from scratch enabling widespread adoption of AI within a few hours.

Personal Spark

What prompted you to pursue a career in Life Sciences? Was there a specific moment in time or influence you can remember? What drives you to work in this space?

There is one particular day that influenced my career in Life Sciences. I was studying physics at Cologne University with a focus on General Relativity and Quantum Field Theory. As part of my course, we had to do practical sessions, which often had the same syllabus since 1970. One day we had a session in the newly founded biophysics department. On the day we were supposed to trap bacteria in an optical laser trap, but the experimental setup was broken. The Ph.D. student that was tasked with our supervision had no idea what to do and just said we should try to fix it and if we fail it’s fine. So, we spent a whole day aligning lasers and heating up bacteria until it worked. From this moment on I fell in love with the life sciences and rarely looked back. Just so much is happening in the field right now, and I feel extremely excited to be part of it.

How did you get your training, if any to be able to build your company?

Two years ago, I only had a very vague idea of what an entrepreneur is and never thought I would be one of them. When we founded DeepMirror in 2019 we were initially simply consulting scientists on projects with no desire to scale up. Then after I finished my Ph.D. in 2021, I joined the Cambridge Accelerate program from the Cambridge Judge Business school. Now I sometimes wish I would have done that earlier because it really changed my perspective and made me leave the academic track to pursue a career in company building. Additionally, I was and am extremely fortunate to have found many great mentors in my circle of friends, two of which I count indispensable in both my personal journey and that of the company.

Can you tell us a little bit about your background & career thus far? What were you doing before you started running a high potential, venture-backed startup?

From high school onwards I was on track for an academic career in physics. As a teenager, I was obsessed with fundamental questions about space and time, determined to make a career in this field. Hence, I enrolled for both a bachelor’s and a master’s degree in physics at Cologne University in Germany. However, during the master’s degree, I became more interested in other fields as well and after a stint in theoretical plasma physics research at Osaka University in Japan I went to Cambridge to study biophysics. At Cambridge, I found myself in a highly interdisciplinary laboratory. My boss was a trained veterinarian turned physicist and my colleagues had degrees in math, biochemistry, developmental biology, and engineering. This environment was intensely stimulating as I routinely had to combine concepts of physics, biology, and computer science to achieve my aims. During my degree, I also familiarized myself with advanced machine learning concepts and started applying them to my work and that of colleagues. Especially, while working with colleagues and collaborators I became passionate about AI as a utility first, i.e. AI that is designed to be used and not simply as an intellectual challenge. After a year of research (mainly in my free time), we developed our own algorithm, now called DeepMirror Spark, that makes AI accessible to our users by reducing the amount of data required for AI training. This is how DeepMirror was born.

Company Overview

What problem is your company solving?

Conventional AI requires large, annotated datasets to learn to extract insights. Annotation (for example marking outlines of tumors in brain images) must be done by humans by hand for each new dataset and repeated periodically and for new applications. In life sciences, this is difficult because datasets are often small and ever-changing. For example, each clinical trial would need a new AI for a different small dataset. Additionally, the process of annotation often requires actual laboratory work done by highly trained individuals which is expensive. These factors mean that adopting a single AI might take months which is not feasible for most applications. Thus, the life science sector is still lagging most industries with the adoption of AI.

How did you become motivated to tackle this particular problem?

During my Ph.D., I began applying machine learning to bioimage data mostly for my own research. Subsequently, we founded DeepMirror to assist others in deploying AI for their purpose. While doing so we noticed that our clients frequently needed AI for specific tasks and then another AI for the next making it difficult to train a single AI for all. In the end, these people often reverted to slow and error-prone manual analysis. It felt like everyone was talking about how great AI is and how I will help the life sciences, but nobody was truly enabling it. For me, this was the moment at which I decided to jump into the fray and bring agile small data AI to the life sciences.

What does your company do?

We enable the full AI cycle from creation over deployment to application for biomedical data.

Now in technical language, what are the specifics of what your company does?

We are creating novel and more efficient ways to train and deploy AI from small datasets. To do so we develop a core set of algorithms that employ a (for now) proprietary mixture of semi-supervised learning, generative adversarial networks, data augmentation, and synthetic data. With this approach, we can reduce the amount of labeled data required by 100 fold while leveraging unlabeled data at the same time. Using both kinds of data allows our users to train AI on partially labeled data so that they can truly hit the ground running.

Why does your solution matter for the world when you get it right?

AI in its current state heralds many unfulfilled promises. Only a new generation of utility-focused AI companies can change that. If we succeed in our vision we will enable the widespread adoption of many small AI helper tools in Life Sciences, and bring about transformational changes in drug discovery, clinical medicine, and foundational research.


What is your company’s founding story? How did you meet your co-founders?

Our genesis was a long process that took place over several years, but it started from a very particular moment. My co-founder Andrea and I just went through the process of preparing a scientific publication for an AI-based neuroscience software we developed together. During that time many people came to us and asked whether we could alter the software for them which we had to turn down because there was no scientific novelty in this kind of customization. Then while working in the laboratory late at night he suddenly turned around to me and asked: “What if there were a company that made it easy for these people to make their own AI?”. DeepMirror was born.

We initially built custom AI for individual clients in our spare time. However, we changed direction during the COVID lockdown in spring 2020. At the time we were facing the fact that clients had had large-ish datasets but struggled to perform large-scale annotation of their data. Thus, we decided to spend time researching algorithms that would help us solve this problem. After two very intense months, we developed DeepMirror Spark— an algorithm that made it possible to both learn from small, annotated data and unannotated data. With this new tool, we then joined the Cambridge Judge Business School Accelerator program to develop our pitch and raised our first round of financing in July 2021.


What are some of the notable milestones your company has achieved thus far?

Personally, I think we achieved three major milestones thus far. The first one was the initial inception and testing of the DeepMirror Spark algorithm. To this day I am immensely proud of this. At the time we had no clue whether our approach would work and there were weeks of despair and short bursts of euphoria.

Our second milestone came with the securing of our first round of funding which happened much quicker than we anticipated this June.

Finally, we just hit our third milestone: We spent the last 5 months building an MVP web platform on which clients can access and use DeepMirror Spark on their own data. This was a massive undertaking as the algorithm had to be rewritten and packaged into a scalable infrastructure. I still cannot believe that we did this in so little time but it is true.

What are some of the biggest hurdles ahead?

In the short term, we need to onboard a few pilot clients to demonstrate product-market fit which becomes a tricky task of integration as everyone uses a different software architecture and has slightly different requirements. We will have to navigate the fine line between staying scalable and helping customers adapt to our technology. In the medium term, it is all about team building. How to assemble a world-class team that takes the vision forward? Even the greatest ideas are worth nothing without brilliant and driven people working on them. Finally, long term we face the question of whether we will be able to also provide small data solutions outside the life sciences sector. In theory, the DeepMirror Spark algorithm applies to any kind of data and has the potential to be vastly improved over time. We hope to outgrow the life sciences sector in a few years and truly push the widespread adoption of AI as a utility.

Pay It Forward

Throughout the journey, what have been some of your biggest takeaways thus far? What advice/words of wisdom would you share from your story for other founders?

Put yourself out there. Initially, I rarely talked about the business and did not engage with any business folk around Cambridge. In hindsight, I would change that and leave my comfort zone much earlier. Also, talk to your friends about your projects, you never know who they talk to.

What are some of the must-haves for an early-stage Life Sciences startup in your eyes?

This is probably highly subjective. For me, it is all about the focus & the team. Yeah, the idea can be amazing and cure aging forever but without a team that is able to become an organization the idea is worthless. So, look around you and be aware that you are basically marrying your co-founder and early employees. Then focus. Try to really really narrow down what you want to do right now and keep questioning whether you are really focused enough. There are many “cool” things one can do, especially as an academic, but in the end, a company needs to build something that makes money and brings value to customers.

What are some of the traits that make a great founder? What type of risk profile/archetype does someone need to have to be a founder in your opinion?

I don’t think there is a particular founder profile; I met very different people that were all successful in their own way. Maybe one thing I noticed that these people had in common was that they were not crushed by the weight of responsibility.

For folks coming out of academia, what advice would you share?

Cool ideas are not necessarily good products. When I came out of academia, I had it ingrained in me to always do cool and novel things that were not necessarily sustainable or serve any other purpose apart from being new. In business, you must take these cool ideas and turn them into products.

Not everyone knows everything. Often founders have to learn either the science or the business side better. What advice would you give for someone picking up a new skill set such as this?

For me this already happened twice in my career: Once when I jumped from Physics into Biology and now again when I jumped from academia into business. In both cases, it helped me a lot to find an environment that was full of experts in those respective fields and then keep asking them lots of (stupid) questions. If you are not feeling out of your depth you are probably not doing much worthwhile.

What advice for managing and hiring a great team can you share?

Set up a hiring process from day one and never micromanage! Truly great people want ownership over their work and when you hover over their shoulders, they feel constantly patronized. To hire these people, it is paramount to set up a hiring process that tests exactly what you need and not just people you get on with. If you want to hire a software engineer, for example, you should only make them do tasks relevant to their actual work. In the end, you want to hire people for a job and not someone with tangential knowledge of many areas.

📣 If you enjoyed this post please clap 👏 & comment 💬 to let us know

❤️ Thanks Sasha Eremina for your help in putting this together :)

Alix Ventures, by way of BIOS Community, is providing this content for general information purposes only. Reference to any specific product or entity does not constitute an endorsement or recommendation by Alix Ventures, BIOS Community, or its affiliates. The views expressed by guests are their own and their appearance on the program does not imply an endorsement of them or any entity they represent. Views and opinions expressed by Alix Ventures employees are those of the employees and do not necessarily reflect the view of Alix Ventures, BIOS Community, affiliates, and content sponsors.

Join BIOS Community 🎉

Become a member, continue the conversation, connect with like-minded Life Science innovators, access exclusive resources, & invite-only events…

Apply to Join — Membership Application

For More Interesting Content 💭

🧬 PodcastStream Full Episodes

🧪 YouTube Watch Videos

🩺 Twitter Explore Feed

🦠 LinkedInRead Posts

Subscribe to BIOS Newsletter for special content: Click Here 🔥



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store