AI has gone so far as to potentially change drug development, and these developments have the potential to make a difference in people’s lives. On this episode of BioTech IQ, Ammon interviews Noam Solomon, the CEO of Immunai. Immunai is developing an AI platform that leverages big data, single cell multiomics, and machine learning to bridge the gap between casual immunology and translational disease biology. Their platform is mapping the human immune system at an unprecedented level of granularity, allowing them to identify the right targets to go after. In this interview, Noam shares his knowledge on reprogramming the immune system for better diagnosis and treatment. He shares his amazing team and how they make patient’s lives easier and meaningful. Tune in to learn more about immunotherapies!

Listen to the podcast here

Mapping The Human Immune System: Using AI To Change Drug Development With Noam Solomon

As you can tell, it has been a while since I produced an episode. It’s interesting because I have a pretty full production schedule with a number of episodes that need to be edited but I’ve taken a break. To give you a little bit of background on myself as I like to do, I started the show a few years ago. My first episode launched on August 13th, 2020. We’ve hit a milestone and I wanted to do a special episode, but I’ve been so busy that I wasn’t able to get to recording that special episode. The summer ended. I do have children. It was their first day of school. Everyone started. Things are shifting.

As I wrapped up the summer with my family and took a little bit of a break, I am back to making sure that I produced these episodes for you. I also want to encourage you that even though I’ve been busy taking a little bit of a break from the show itself, I have put some time into producing some videos that are on the YouTube channel of the show. You can go watch some new videos that I’ve been producing there with some onsite interviews with different staff of different companies and some of the leadership of the companies. I’m also getting an on-site tour of the companies and then putting that into a video for you to watch.

Let’s jump into who our guest is. Our guest on the show is Noam Solomon, the Cofounder and CEO of Immunai. Noam has a background in computer science and mathematics. What Immunai is doing is they’re leveraging big data, single-cell multiomics, and machine learning to bridge the gap between what they call casual immunology and translational disease biology. Their platform is mapping the human immune system at an unprecedented level of granularity. This is allowing them to identify the right targets to go after.

We have had this discussion before on the show about what is it that AI can do. I don’t want to say gold, but what we know it can do is help to accelerate the discovery process in the early stages. In the world of lean startups, I’m not talking about the biotech world, there’s this concept of fail fast. The faster you fail, the more money you save and the sooner you realize that your idea isn’t going to go anywhere where you can pivot to the right idea faster. AI has the potential or the ability to help us do that in the biotech world in developing therapeutics that change human life.

Immunai is on that track. They’re partnering with biopharmaceutical companies as well as top institutional research organizations that are all involved in helping in this process. Without further ado, I will let you get to the show. Thank you so much for being here again. Some of you have reached out to me. I put an ad in my show episodes that I don’t have any direct sponsors yet. I’ve spoken to a couple of different companies. I’m getting some ideas from some of the listeners of the show that have reached out to me. Shout out to those of you that have done that. I appreciate it. Thanks for reaching out. I love connecting, talking and hearing from you. Please continue to do so.

Welcome to the show, Noam Solomon.

Thanks, Ammon. It’s nice to be here.

It’s good to have you on. You and I were commenting before the show got started a little bit about the long process to get you on here. I’m glad we got you on. I was doing some research on Immunai. There are some awesome things going on. I’m glad to have you here.

Thanks. It’s great to be here.

Let’s go ahead and jump right into this conversation. First of all, as was noted in the intro, there are a couple of interesting points about your background. You have a PhD in Computer Science. Later after starting your career, you went back and finished a PhD in Mathematics.

It’s the opposite.

I apologize. I got it mixed around. First of all, congrats on the double PhD. That takes a lot of dedication. I’m curious. What was it in your early life that put you in this direction to study mathematics and computer science to get you on the path to where you’re at now?

I’ve always liked puzzles and challenges ever since I was very young. In pure math, you can ask for something more open-ended. There are so many interesting questions to ask. I was exposed to it from an early age. I also got a lot of encouragement from my family, my mom, my grandfather, my dad, and others. Looking back, solving problems is not only math but at an early age, I think it was mostly pure mathematics.

How do you study immunotherapies? You study the immune system. How do you study the immune system? That's a big question. Share on X

That would make a lot of sense given what I know about computer science and mathematics. You’re looking to solve a riddle in a sense. With math, you follow the process. I hear that all the time although I’ve always struggled with it. If it’s so easy to follow the process, then how can I follow the process and come up with the wrong answer?

There are multiple ways to look at mathematics. I don’t necessarily think about it as a deterministic process leading to one answer. The other versions of how you think about research are much more open-ended. Mathematicians can figure out how to start with a set of axioms and rules of deduction, and then build wonderful new theories. Some of these theories are leading to concrete equations and things that are like what you’re describing but others are much more open-ended. As a mathematician, especially if you’re a mathematician, I always thought about math as a similar area of research to philosophy and other things that people commonly recognize as more open-ended. Others view mathematics as solving questions and doing some exercises to pay bills. It’s not just arithmetic.

The type of math you’re describing is beyond what we learn in the early stages of school or education. You understand this basic process. You’re talking about figuring things out that you don’t know exist yet.

I’ll be concrete. One of the problems that influenced me from an early age was called Fermat’s Last Theorem. Somebody named Pierre de Fermat posed the question almost 400 years ago. He thought he had a quick and nice solution. He wrote it on the back of a notebook. Many mathematicians for centuries tried to solve this question. Only in the ’90s or the 20th century that it was solved, but the solution wasn’t a few lines of code. It was a couple of hundreds of hard mathematics. The person that solved this problem, Andrew Wiles, is somewhat of a hero to me because he worked on this for at least seven years. He was exposed to this question when he was a child and he never gave up.

It’s never giving up on something that is quite mysterious. I don’t know exactly how it’s solved. Even though the statement of the question is fairly simple. It’s something that you can explain to a child. I can explain to the audience what the problem meant but maybe it’s going to take us too far away. What drove me to this question was a very simply articulated number theory question that everybody can understand within a minute. For centuries, nobody was able to tackle this question except for this one person. His name is Andrew Wiles. He decided to spend his career at Princeton to solve this question. This type of journey that people are taking is inspiring to me.

How do you go then from what you studied throughout your academic career to then going into technology? I’m looking over your LinkedIn profile here and some of the things that you did. You moved through your career, and then here you are at Immunai. It’s looking at AI. This is a hot topic. There’s a lot of discussion on AI in the biotech industry and how this is going to help discover things in the earlier stages so that we don’t spend so much time and money, and then have a failure in the later stages. How is it that you go from what you did throughout your career thus far to interacting with the biotech industry?

It’s by chance. I didn’t plan for it to happen. My journey was inspired by questions and challenges that I thought were interesting. A few years back when we founded the company, I was a postdoctoral fellow at MIT in the Pure Math department after spending a year at Harvard CMSA. Through a friendship with my co-founder and CTO, I decided to think about what could be a very interesting way to leverage our respective skills. We both came from similar backgrounds in math, computer science, and ML.

BTI 55 | Drug Development

Drug Development: Building the capabilities to do single-cell immune profiling with the different cells and reducing the batch effects and all the noise that comes with the territory is both a challenge on the technology side and the immunology side but also on the computational side.

 

One of the topics was computational biology. We quickly landed on what is being done today with therapeutics. We started looking into immunotherapies, which are specific therapies for the immune system, especially for different types of cancer indications. It was an interesting story. My co-founder and CTO’s name is Luis Voloch. His grandfather had cancer at the time. For him, it was a personal question. It was interesting to me why his grandfather was going through what seemed initially successful therapy. The tumor was regressing. The adverse events were too difficult for him. He decided to stop the medication.

I landed on a problem that not only was interesting from a data science perspective but was also important enough to change someone’s life. Initially, we started researching the question from a few different angles, “How can we find relevant data? How can we look for what relevant researchers published recently?” Pretty quickly we understood that there is a very important problem for us to try and solve, not only for one person or another. This led to the genesis of Immunai.

Immunotherapies have been also a hot topic in and of themselves aside from the AI conversation. I had been in recruiting for over seven years as a professional recruiter. I worked mostly with medical. A few years ago, I got into the biotech industry and was having these conversations to understand how the clinical trial process works and all that. I came across a couple of clients that were working on immunotherapies. I tried to at least understand the basics of it. I did a little research on their websites. I was like, “They’re training the immune system how to fight cancer and do something it doesn’t know how to do.”

It is revolutionary. One thing I wasn’t aware of though is the challenges that come with developing that kind of therapy. It sounds like your CTO and co-founder’s grandfather experienced that. You were having this benefit watching the tumor regress but then at the same time, the adverse events he was experiencing were too much to deal with. We talked about the genesis already of Immunai in what you were sharing. What is it that Immunai is looking to do to fix that? How do you see your platform being able to tackle that problem that you explained or at least help tackle it?

When we started thinking about the problem, we had something that was both an advantage and a disadvantage. None of us came from this space. We’re not medical doctors, immunologists, or pathologists. We started interviewing people. We started at MIT. I was sitting there. We interviewed a few people from MIT. Within a few weeks, we landed with someone that became our first scientific founder. His name is Ansuman Satpathy. He’s a professor at Stanford.

He introduced us also to Danny Wells. He also became a scientific founder at Berkeley and then Parker Institute for Cancer Immunotherapy. I also joined Dan Littman from NYU. We had this group of five people. Luis and I were the co-founders. Ansu, Danny, and Dan became the scientific founders. They have a rich background in pathology, immunology, and computational biology. They have to think through the problem in terms of what is the right data to look into.

Immune therapies are therapies for the immune system. They are triggering the immune system to be better. How do you study immunotherapies? You study the immune system. How do you study the immune system? That’s a big question. There are many ways you can study the human immune system. One person that I respect told me, “It’s infinitely complex.”

The platform that Immunai offers can and should be used for the acceleration of the development of therapeutics. Share on X

There are so many different organs involved and trillions of immune cells, cytokines, and ligands. Everything is complicated in the human immune system. The basic object of cells is interacting with one another. They’re being deployed from peripheral blood into the organs. Being able to measure the human immune system is something that was thought of as prohibitively hard. That’s what we were trying to solve from the beginning.

Immunai was set to fully map the human immune system and later reprogram it with different types of therapeutic modalities. The initial goal was, “Let’s try to figure out a way to leverage immunology, pathology, and single-cell technologies together with big data and engineering so that we will be able to build the largest database in the world, and then use machine learning to mine insights from this database.” This was the real story of what Immunai brought together to bring forward some contribution in this space.

Help us understand what makes the Immunai AI platform unique from other platforms that are out there. I’ll leave the question at that. I was going to give you some background on why I’m asking that question but we can leave it at that.

First of all, we are leveraging different scientific disciplines in a way that was not brought together. One of them is the way that you measure cells. We leverage what they call multiomic single-cell sequencing technologies. Maybe I’ll try to be high-level and provide some details. We want to measure cells. The cells come from different parts of the body, for example, the blood. They also come from solid tissues and bone marrow. When you measure tissues, you can decide how granular you want to measure those tissues.

We leverage single-cell technologies that are giving you a resolution of being able to measure every single cell separately. Within every single cell, we are measuring mRNA, surface proteins, epigenome, and TCR and BCR sequencing. We even measure CRISPR because we do functional genomics and also do spatial profiling. We’re trying to look at a cell from a 360 point of view perspective, which is complicated.

Building the capabilities to do single-cell immune profiling with the different cells and reducing the batch effects and all the noise that comes with the territory is both a challenge on the technology side and the immunology side, but also a big challenge on the computational side. Immunai is unique in that we are bringing those technologies together. We’re also building the largest of its kind single-cell immune atlas, which we call AMICA. We have big data and ML capabilities that allow us to mine data from such a large database.

When we’re talking about the database, every entry in the database is a terabyte of information. We have many thousands of patient samples. We’re going to get tens of thousands and then try to grow even more. That means that you do complicated computations on a large database. That requires certain engineering capabilities. In this space, we shouldn’t underestimate the importance of not only building machine learning or AI but rather infrastructure and engineering capabilities. These are the areas that Immunai has uniquely differentiated.

BTI 55 | Drug Development

Drug Development: This technology or platform is capable of doing more than either discovering new therapeutics or accelerating new therapeutics. In this sense, Immunai is leveraging its profiling platform to help our partners with the problem that they have.

 

You mentioned the engineering capabilities. I saw on your website that you talk about engineering-first. I’m curious how you define success. What’s the target here? It’s easy to say, “We can continue to develop, come up with new ways of looking at things, and continually progress,” but how are you defining Immunai’s success at this point?

I’ll give you two answers. The first is more high-level. Immunai has a vertically integrated platform that is evaluating therapeutics. Success for us is being able with our partners that are both biopharma and biotechnology companies to better develop and accelerate the development of therapeutics. Success is having patients that are receiving successful therapies. They’re going to do better and improve patient outcomes. That is what success is to me. I can also give you a more detailed answer because I didn’t say what the platform is capable of doing and how we measure success from a platform perspective. The high-level perspective is improving patient outcomes by better delivering therapeutics.

Let’s hear the more detailed version. Most of the people in our audience are in the biotech industry. We have very experienced executives, scientists, physicians, and clinical operations. Some are higher on the scale of scientific knowledge. Some are a little lower but all generally understand what’s happening in the industry.

The more detailed version is that we are mapping and profiling the human immune response to different types of therapies. We are mapping patients longitudinally. We are taking a few different data points, especially in pre and post-therapies. We are looking for what we call the difference or the delta in the immune response pre and post-therapies, not for one patient but as many patients as we could.

We’re doing it for standard-of-care therapies, clinical trials, and experimental medications. Over time, we are able to unlock the mechanisms of action and resistance to therapies, why drugs don’t work for certain patients and they are working for others. From there, we are deducing the explanation that they’re helping us identify novel biomarkers and also targets for novel therapies.

We are taking those targets for novel therapies. Together usually with partners, we are looking at the next generation of therapies. The way that this becomes very interesting for also the engineers and the machine learning experts is that you can find those patterns in qualitative data, but those qualitative insights are coming with certain statistical significance. You then come up with insights.

You think that this gene in this immune cell type is highly correlated with resistance, which is fine but the next layer is being able to apply functional genomic technologies to functionally validate that this insight that you have in a qualitative manner you can validate in a causative manner. Going from correlation to causation is what we leverage with our functional genomic layer. This helps our partners gain more confidence in the insights that they have and the drug candidates that they develop.

Being able to fail fast is a good recipe for succeeding better. Share on X

You’re helping your biotech and pharma partners understand the causation of something. They’re able to move forward from there with whatever treatment or therapeutic that they’re developing. They’re able to more closely match that therapy to the indication that they’re treating. Am I understanding that correctly?

Right. I was a little bit vague because of a few different use cases. One of them can be, “This is a drug candidate. Did we choose the right indication for this drug candidate? Did we choose the right combination therapy?” Oftentimes the therapy that the patient is receiving is not one monotherapy but a combination of a few different therapies together. Did we do the right patient segmentation for the clinical trial? All these questions are relevant for therapeutic development but we can also go earlier and say, “We found something that looks like a good target for a novel therapy.”

We will also do novel therapy prioritization from target combination, target validation, and drug and compound verification. We have a suite of capabilities. This technology or platform is capable of doing more than either discovering new therapeutics or accelerating new therapeutics. In this sense, Immunai is leveraging its Immune profiling platform to help our partners with the problem that they have.

I want to be careful here. There are certain things that a company like yourself and at the stage that you’re at are able to comment on. I preface that question with that because you’ve mentioned our partners. Are you directly partnered with any therapeutic development companies at this time that are working in this space?

Yes. Because we care about the privacy of our partners, we don’t share names but we are working with both Big Pharma Fortune 500 and 100 companies. We also work with smaller biotechnology companies. Our main business model is to accelerate the development and discover of new therapeutics with our partners.

The end goal of what you’re doing at Immunai is to license out your technology to companies that want to access it and use it for their development. What’s the end game for you with this?

The endgame is that we believe that the platform is able to unlock certain resistance questions. We want to use the platform to improve patient outcomes. We want to use this platform to accelerate the development of therapeutics with our partners. In terms of licensing, our licensing still depends. We haven’t decided whether we’re going to license the target that we have developed or we’re going to co-develop with others. It depends on the right partnership.

BTI 55 | Drug Development

Drug Development: When you have a new immunotherapy candidate, you want to know whether the immunotherapy is going to work in patients. Unfortunately, none of the existing methods are incredibly accurate.

 

I appreciate you gracefully accepting those questions. I don’t want to pry too much but we’re on the show. We want to talk a little about what the plans are for the future so people can understand that.

I would be happy to answer.

This is the final question I wanted to ask on what you’ve got going on with Immunai. You’ve touched on this already a little bit but it’s a direct question to understand. How do you foresee the overall application of your technology in the marketplace moving forward in the future?

We believe that the platform and the offering that we have can and should be used for the acceleration of the development of therapeutics. We want to do it because it’s one company and one platform. We want to do it with the partners we believe are the most innovative and the most cutting-edge and their ability to leverage such capabilities and move faster. I’ll give an example.

When you have a new immunotherapy candidate, you want to know whether the immunotherapy is going to work in patients. Unfortunately, all existing methods are not incredibly accurate. Lab methods have models. In vivo models in animals are not going to recapitulate human biology. One of the statements that were told to me a few years ago is that we can cure every type of cancer in mice.

We need to bridge the gap and Immunai is working on recapitulation of certain immune ingredients that the drug or the immunotherapy is supposed to trigger in the responder to the therapy. What we are doing is trying to test whether those immune ingredients or these immune reactions are happening. If it doesn’t happen, it doesn’t necessarily mean that the drug is bad or good. It means that the immune conjecture that we have was not the right one.

The thing is when you develop therapeutics, you need to have a clear objective and a clear conjecture. We help unlock whether the immune therapy is activating the right parts of the immune system. It’s a cascade of events. These accelerate the process. In terms of use cases, we want to leverage this platform’s capabilities to tell our preferred partners, “You can take it and accelerate the process of clinical trials remarkably.”

Study biology and medicine. It's going to really make your life more meaningful. Share on X

“You can improve the way that the drug will achieve the right immune response. You can better select the indications. You can better do the segmentation of patients. You can even optimize for the right dose escalation and the right combination therapy.” All these things will be delivered with the Immunai platform. This is in the next few years the main market business model we have.

That’s exciting. I talk to people all the time in my personal life as well as business life. Here’s one of the things I always hear from people in the general public. I’ll use the term non-biotech community people. You hear, “I know that there’s a cure for cancer that exists. They are hiding it from us.” When people start to talk like that, because I run the podcast and work directly in the industry, I usually say, “I want to respectfully disagree with you on that. Here’s why.” We will then have that conversation.

One of the things I always tell everybody is based on what I can tell from the companies that are in early-stage development or even mid-stage development and even a couple that are late-stage things like Kite Pharma and things like that, “What is around the corner is what you’re talking about. It’s this cure that we’re seeking,” but the general public tends to misunderstand how complex cancer is and how complex it can be, depending where it’s at in the body.

It’s not necessarily the cancer itself but being complex where it’s at in the body. It’s how the treatments that we have affect the human body. This may kill cancer but when you put it inside of a person and their biology reacts with that, it’s different. We can cure all cancers in mice but that’s not a person or a human. We’re trying to figure it out. That’s the other thing. People say, “The FDA, the bureaucracy, and all that.”

I do agree to a certain point when it comes to the point they’re trying to make. What I’m trying to help people understand is that the reason the process for developing medicine is so long is that at every step of the way, there have to be checks and balances and testing going on to make sure of two things. Number one is that it’s safe for people to use. There’s an acceptable amount of risk, which we consider safe. The final thing is it has to deliver a significant benefit above what already exists in the market.

If a medicine cannot do that or if it can deliver a significant benefit above but not do it safely, then the FDA or whatever regulatory authority agency in the world is out there is not going to grant those approvals. You may read on the internet or hear someone talk about an exciting new therapy, but that doesn’t mean it’s going to deliver that safely for people to be able to consume or use. That’s exciting what Immunai is doing. Hopefully, it’s helping to shorten that a little bit.

The FDA is responsible for the safety and efficacy of the drugs that we end up giving to patients. A drug may be efficacious but it’s also not safe. Those types of use cases can benefit from having a platform to personalize or understand better the reasons why certain therapies, molecules or compounds are going to trigger a toxic immune response.

BTI 55 | Drug Development

Drug Development: The immune system is complex. If we are able to identify the early network effect that is supposed to happen when you are giving this immune agent, and you can check in vitro in the lab that this immune response is not elicited by the drug, you can kill the clinical trial.

 

Being able to identify or stratify this beforehand can improve the results of the clinical trials because, at the end of the day, you want to find the right therapy for the right patient. Being able to understand or look under the hood helps it. This was the motivating reason for founding Immunai. It’s to be able to look under the hood. In this case, the hood is the immune system because the immune system is responsible for the good and the bad immune response that may lead to a bad physiological response.

That’s exciting. The most important thing is the patient as you’ve talked about. I also see from an investor standpoint this being a tool. I don’t remember what the exact percentages are but it’s something like 7 out of 10 drugs fail. It’s 30% to 20%. I’m not putting the exact numbers. I talk to people. 7 or 8 out of 10 drugs or something like that fail. Through some type of platform such as Immunai, what if we could only get things into the process that we have a better understanding that they will make it through to the end? It’s going to be safe and efficacious.

First, I want to be careful in the way that I articulate myself. I completely agree that there is a process. The process is lengthy but it’s also not data-driven enough. As I alluded to earlier, when you have a compound or an immune engager, that is supposed to engage your immune system and trigger an immune response. What we do at Immunai is we’re trying to map this domino network of things that are supposed to happen.

The immune system is very complex. Some would even say it’s infinitely complex. It’s not infinite but it’s complex. If we are able to identify the early network effect that is supposed to happen when you are giving this immune agent, and you can check in vitro in the lab that this immune response is not elicited by the drug, you can kill the clinical trial. One of the ways to improve the statistics that you’ve mentioned is by being able to kill clinical trials that don’t deal with the right immune response.

Being able to fail fast is a good recipe for succeeding better because those things are not failing 7, 8, or even 9 and 9.5 times out of 10. It’s always taking a very long time, many years, or more than a decade. Sometimes it’s even $2.5 billion to put it. If you can realize earlier that this is not the right therapy to pursue or this is only the right therapy for this indication and this patient stratification, you will be able to save a lot of time and money. The process can be accelerated and improved.

This isn’t necessarily therapeutics-related but what you touched on about failing fast is not a bad thing if the failure is due to discovering that it’s not going to work. It has been around for a long time but there’s a concept in the business startup called the Lean Startup method. You’re probably familiar with it. It says, “I have an idea. I need to go out and validate that idea. Once the idea is validated, then I can move to develop a fully functioning MVP or Minimum Viable Product that I can then start taking to the market.

What most businesses and a lot of people that have a business idea do is they get the idea and go, “I have this great idea.” They spend a bunch of time and money building out that idea. They take it to the market, and nobody wants it. They haven’t validated the business model yet, “How do I deliver that?” I love that. It’s the same concept, which is, “Can we do that in the world of therapeutics, develop better treatments, save time and money, and in the end accelerate the whole process?” You mentioned something though when you were commenting. I wanted to zero in on that real quick here before we move to the next segment of the show. You mentioned that it’s not data-driven enough. What do you mean by that in the process of developing therapeutics?

When you don't know something, you also have an advantage because you're not going to be dogmatic. You're not going to think like everybody else who has been trained in a certain way. Share on X

There is a revolution going on now, bringing more data into making the decisions. From the discovery of targets to the validation of targets for the clinical trial process, it is not data-driven enough. It’s not personalized enough and not validated enough. It’s still being done in a quite trial and error way. One of the advantages of bringing an engineering-first approach is trying to build it to be more data-driven by definition.

For example, it’s not a guess but let’s say that we even have a guess. We found the gene that we think can be an amazing target for a new therapy. We are going to apply functional genomic technologies to validate it in multiple ways. By the end of the process, we are going to be much more data-driven because we have tried different strategies.

One of the things that I love about science is that you start with a hypothesis, and then you’re going to use all of your intellectual capabilities and all of your money to try and destroy it. If you haven’t destroyed it, it probably means that it’s more likely to be right. We are leveraging a lot of our capabilities and technologies to try and destroy therapeutic programs.

When I’m talking about failing fast, we are trying to put a lot of pressure on that hypothesis. If we can, it probably means it’s a better drug. We’re putting a lot of data. We have data coming from hundreds of thousands of patients’ sample data that we did molecular profiling for. We’re trying to do in silico analysis of these things.

At the end of the day, we’re going to put data into the machine that will tell us what the machine thinks is the right way to go. They’re not to take it for, “This is the answer.” We’re going to do a functional validation of this, and then in vitro and in vivo validation to further support it. If we weren’t able to support the hypothesis, that’s it.

The Lean Startup methodology is where high-tech and biotech merge. This is a core part of how we are as a startup. Immunai is more or less 50/50 high-tech and biotech together. We are trying to follow the Lean Startup methodology and figure out better ways to pressure test our assumptions as early as we can. We try to fail fast so that we can move to the right decision quicker than others.

I took a three-day startup ignition course here in Utah. They call this area up and down the 15 Freeway or the highway. They call it the Silicon Slope because there’s so much tech here and a lot of technology companies. There are a lot of people very interested in software development and things like that going on. Interestingly, there’s a small biotech community as well. In the building that I’m in, I met the founder of CancerVAX. It’s a small biotech company here working on cancer oncology as well. I was planning to have him on the show.

BTI 55 | Drug Development

Drug Development: One of the advantages of bringing an engineering-first approach is trying to build it to be more data-driven by definition.

 

Where I was getting to was that I took this three-day course. It’s taught by this guy named John Richards who was a very successful tech entrepreneur from the ’90s. He was like, “You need to get The Lean Startup manual. You need to study this. You need to live it like religion when it comes to your business.” We went through this three-day course where he explained everything in failing fast and business validation. Maybe it’s a little different when it comes to the therapeutic world but maybe there’s a pivot.

You see that happening in the therapeutic world as well where there’s a certain treatment or drug that was going through. They discover it was mediocre in delivering treatment for this but it was fantastic at delivering treatment for X. They’re able to pivot that. The earlier we can figure these things out, the better off everyone will be.

I’m glad to have you on the show talking about these interesting conversations. I’m learning from you and what you’re doing. I’ll go back to this unedited episode before my team gets to edit it and put it out there. I try to take notes and learn from the conversations I’m having with people like you. Thanks for coming on again.

I enjoyed them. The feeling is mutual.

Let’s jump to the last section here. I want to be conscious of your time. I know we scheduled some time here. The audience is devoting some time to hearing you talk. I’m here to facilitate that. I like to ask these final three questions to the guests that I have on. It helps us learn a little bit about your background, some insight into how you think, and also some advice. The first question is this. You look like a pretty young guy. If you could go back to the start of your career and give yourself some advice, what would that be?

There are many things. Probably the first advice is to study biology. I was born and raised in Israel. Since I was 9 or 10, I knew I was going to be a computer scientist or mathematician. I was always into this type of work. A few years ago, I started looking into genetics, biology and medicine. I wish someone told me, “You should look into biology and medicine. It’s going to make your life more meaningful.” That’s something I would have told myself. I’m not sure I would have listened. I was a very stubborn young boy. My dad tried to get me to listen to him.

That’s probably part of the reason why you are where you are though. That’s interesting. You’re finding yourself in this biology and computer science intersection. That’s interesting. Here’s a quick question. What do you do to study biology? When most people hear that, they think, “I need to go to school and get a degree in X thing.” That’s not a bad thing, but in terms of you being a busy startup entrepreneur and having multiple PhDs already and a company, how do you educate yourself on this?

BTI 55 | Drug Development

Fermat’s Last Theorem

First of all, I’m very fortunate to have an amazing team. A lot of my time goes to studying biology both within my work as the CEO of the organization. I also spend time after hours on the weekends because this is a fascinating topic. Immunology is similar in many ways to math. Most immunologists that I spoke with were good in math when they were in high school. A lot of the thinking process is similar.

Maybe the PhDs and my academic training allowed me to quickly penetrate a new subject domain. In the beginning, I had a private tutor. She was a PhD student in immunology. We spent a few hours together every week. I try to go to every journal club. I’m reading books and educating myself as much as possible. A month after we started, my co-founder came to me and my scientific co-founder. I remember they said, “If you want to manage us, you need to study immunology.”

I’m very grateful for this advice because I’m a much better manager now and CEO of a biotech company because I understand these topics. I’m not an expert. Don’t put me in a competition with a PhD immunologist around these questions but I can know my way around these topics. I’m dangerous enough with the level of knowledge that they have. Being curious and always asking questions is the number one thing I try to bring. Maybe it’s an advantage. Maybe it’s a disadvantage.

I came to be the CEO of the company without any background in business, biology and medicine. In the first two years, I asked more questions than anyone else all the time. I ask a lot of questions because that’s who I am. Over time, I learned that when you don’t know something, you also have an advantage because you’re not going to be dogmatic. You’re not going to think like everybody else that already has been trained in a certain way.

For years, they thought that this has to be true. You don’t know that this has to be true because nobody taught you that. That is also an advantage. Immunai has about 150 people. Half of us are immunologists. We always communicate. We have every week journal clubs and scientific discussions. You learn through the job.

That’s exciting. The next question is this. Are there any books or a book in particular that you read that had a great impact on your life, business or otherwise?

There are many. You mentioned one of them, The Lean Startup. I wanted to go back to something I said earlier because it had a tremendous impact on my life. There is a book by Simon Singh. It’s called Fermat’s Last Theorem. Going back to this thing that I mentioned in the beginning when I was in my late teens, I read this book telling the story of Andrew Wiles who later won the Fields medal for proving this theorem.

What is most fascinating about the book is the story of someone that never gives up. He was doing this secretly and not telling anyone even his peers and colleagues because he was afraid that people are going to think he is crazy. This is a problem that nobody was able to solve for centuries. This influenced me to choose my first PhD in Pure Math.

I did something similar in the topic but it also taught me to be stubborn and to believe even when nobody else believes in what you think is right. Even though I read the book in my late teens, I was always stubborn. It was good to see other people that are stubborn and very successful. He’s Pierre de Fermat. I was also reading about him on Wikipedia. The book and name are after his theorem, which is named Fermat’s Last Theorem.

Here’s the last question I have for you. Perhaps we have already been touching on this. Where do you think the industry of biotech is headed in general?

Ten or fifteen years from now, we are going to see a very important change in the way that the biopharma industry is going to be influenced by engineering. It’s also not fair to call it AI. Engineering-first companies in this space need to come to the forefront. Being able to leverage big data, engineering capabilities, and infrastructure to solve complicated data-driven problems like drug discovery, development and manufacturing.

We are going to see a major disruption of the previous way that drugs were discovered, developed, and manufactured. We’re going to see hopefully changes in their regulatory process. We’re going to see more open-mindedness to the personalization of medicine. People need to be patient because it’s not going to take a year or two. Maybe it’s going to take twenty years, but it’s the direction of the industry.

With that, I appreciate you coming to the show. Thanks for dedicating the time. This has been awesome. It’s cool to understand what’s happening with AI. I’ll say AI even though you said it’s not necessarily fair to only say AI. It’s cool to see what’s happening with your company. I look forward to seeing how things turn out in the next few years.

Thank you very much. It was a pleasure to spend time with you.

 

Important Links

 

About Noam Solomon

BTI 55 | Drug DevelopmentNoam Soloman is the CEO and co-founder at Immunai, the first and only company to map the entire immune system for better detection, diagnosis, and treatment of disease. Leveraging single-cell technologies and machine learning algorithms, Immunai has mapped out thousands of immune cells and their functions, building the largest proprietary data setin the world for clinical immunological data. Prior to co-founding Immunai, Noam had a dual career in both the industry and academia.

Noam is a double Ph.D., and served as a post-doctoral researcher in the Mathematics department at MIT, and in the center of mathematical sciences and applications at Harvard University. In his research, he developed and applied tools from algebra and Algebraic Geometry in the study of classical problems in combinatorics. Noam also worked as an algorithms developer in the Israeli defense forces and subsequently as a data scientist, consultant and head of data science in several hi-tech companies in Israel.