Maxim's 2016 and 2017

This is a yearly update I send out to close friends and mentors. This installment is from January 2018. (Previous updates here.)

Maxim's 2016 and 2017

Over the last year and a half, I graduated from college, dove further into data science and biomedical research, and moved to New York for a gap year before I head to grad school. I’ll share stories from my work at a big New York City hospital, from my improvized travels on the cheap, and from my recent side projects. Links to photos and music from 2016 and 2017 are at the bottom.

My last year at Princeton

In June, I completed my undergrad at Princeton. That’s right. This time around, I actually graduated. I don’t have the best track record here… Four years ago, I left high school early without a diploma (Princeton took a GED). Earlier, I skipped the last year of middle school. And I’m not positive I was in town for my elementary school graduation. So college graduation was a big deal, at least for my parents.

(And I finally got a high school diploma in the mail. Apparently, some of my Princeton classes fulfilled their requirements.)

I graduated

My four years at Princeton were hands down the best of my life so far. Though the freedom of adult life is intoxicating, I miss Princeton madly — the daily routine of captivating seminars, encountering friends from all circles of life, the nightly political roundtable over dinner…

The nightly roundtable at Princeton

Besides solidifying my intimate circle of dear friends over the last year, I enjoyed some classes out of left field: the political history of health care reform (I was surprised by the ACA’s clever “hacks” to align incentives in the insurance market, though I suppose that’s irrelevant now); ecstatic dance (an opportunity, hidden deep in Princeton’s course catalog, to spend two hours lying on the floor one day, and slither around in the grass and mud imitating a worm the next — plus a fantastic instructor and a tabla player who enjoyed jamming with students); and health and science journalism.

The journalism class has actually been on my mind a lot recently. I’m trying to publish a story I wrote for that class about how software-designed cancer vaccines upset the established paradigms of medicine. To my knowledge, these drugs mark the first time artificial intelligence has played such a direct role in drug development. I’ll send an update when this story about the challenges of regulating A.I. in medicine is available. I didn’t expect I’d enjoy journalism, but I fell into my old practice of calling people whose work I found fascinating to ask how they got there, how they saw the world — with the extra step of putting their stories, and the complicated science behind them, to paper. Now that’s fun.

I’ve also pursued my growing interest in biomedicine. My Princeton senior thesis extended the research I worked on in summer 2016, which I spent in New York City at Mount Sinai’s school of medicine in Jeff Hammerbacher’s group. You could say Jeff has an atypical perspective on academia and biomedicine, having just taken his data science company public and being a professor without a PhD. I wanted to emulate his transition to biomedicine and to understand his way of thinking as a scientist and investor.

We focused on measuring the number of immune cells of different types in the “microenvironment” that surrounds a tumor — a problem that quickly got me into the weeds of both genetics and machine learning. Here’s the gist:

As part of the immune system’s response to cancer, immune cells often infiltrate the region around a tumor. Figuring out the contents of this area could give some important clues about the state of the cancer. Perhaps it could even explain why impressive new immunotherapies only work for a fraction of patients. If we knew what makes those patients unique, that would do a lot of good for targeting those treatments, maybe even developing better ones.

With Jeff and several colleagues who have taught me immensely, I’ve built a new method with the flexibility to characterize more cell types than possible with earlier manual measurement or computational approaches. My colleagues and I presented the work at a UPenn symposium in May, at the International Cancer Immunotherapy Conference in Germany in September, and at the Society for Immunotherapy of Cancer meeting in D.C. in November. Here’s the preprint — a complete manuscript is on the way.

Working at Mount Sinai Hospital

The next chapter: a PhD at Stanford bridging machine learning and immunology…

In recent updates, I recounted my transition from software engineering to data science, shared my excitement of solving business challenges by translating them into statistics questions, then integrating the results back into the business, and told of my forays into building digital products in my spare time.

Recently, I committed to channeling these interests into biomedical research, which I’ve been diving into over the last year and a half. This isn’t a standard choice for a computer scientist these days. Let me explain why I decided research is the right next step.

What to study?

My current focus is to apply machine learning to problems in biomedicine. I believe computer science can create a paradigm shift for a number of biomedical challenges. I’d like to gain a deeper mastery of machine learning methods while simultaneously contributing to immunology research. Long term, I’d like to build products and experiences at the intersection of science and technology and expand my impact beyond producing knowledge and writing papers.

Though cancer immunotherapy is a hot topic in immunology, I believe there’s more room within the field for data science-powered improvements. And when exposed to the constraints and traditions of medicine, AI needs to be developed and productionalized much more carefully than we are used to in the tech industry — as reinforced by what I learned when researching my forthcoming article on algorithmic cancer vaccines. This makes me highly interested in the systems side of machine learning, as well.

Why grad school?

Pursuing biotech seriously requires far more education to make a meaningful contribution. A PhD program provides a protected time to learn the intricacies of immunology, find useful problems worth attacking, and make progress.

And I thrive when surrounded by fascinating people with similar interests. I feel a desire to keep learning. Why stop now?

Grad school will develop my craft as a researcher while I clarify how to apply it. In the long term, I can see myself extrapolating from my current interests and also tackling problems related to health care delivery systems and to bioethics. There’s a significant gap between the lab and the clinic that lends itself to innovation. My background in software engineering and product development makes me interested not only in exploring the fundamental scientific questions, but also in making practical changes.

But it will take significant immersion to get the knowledge and credentials to make an impact in biotech. So I’ve been proactive about studying the landscape of biomedicine and healthcare. One goal for the upcoming year is to consolidate my reading and conversations into some notes — indeed, a tech-VC friend suggested I write a biotech market map to build mental models for the landscape. I see the piece about the implications of bringing software to biomedicine as the beginning; digging into more stories like this will help me get the bearings of my newly adopted field.

Where to study?

First, I applied for the Rhodes scholarship to study statistics and health policy at Oxford. A friend of mine had just won the scholarship and convinced me to apply. It seemed like a lottery to join an incredible intellectual community for two years. Why not throw my name into the hat? And I believed I’d be equipped to contribute meaningfully at the intersection of fields that define biomedicine and healthcare when I returned to the U.S. to pursue a PhD.

Though writing the application was a ton of fun, the Rhodes didn’t pan out. Thanks to the extraordinary support of eight referees, I was endorsed by Princeton but didn’t make it to the final stage of the process.

I cast a narrow net in my U.S. grad school applications, searching for professors I’d fit well with and only applying to programs I couldn’t turn down. But that backfired! I faced a heart-wrenching choice among seven schools: Stanford, MIT, Columbia, Berkeley, UCSD, UW, and Princeton.

Visiting each school over the course of a month to breathe in its culture was exhausting, but exhilarating: a nonstop science party across the U.S. on someone else’s dime!

I settled on Stanford, where I meshed very well with professors and peers. I fell in love with the campus environment — though I confess, I’m still concerned about moving to a town populated almost exclusively by tech people. With three years of funding from NSF, I’ll have substantial freedom to continue my immunology and machine learning research.

… But first, a year off working on biotech + medtech in NYC

I’m now back in NYC’s East Village, with a view right onto the river. Life here is vibrant. The density of interesting people in New York is unbeatable, and this is a jazz nerd’s mecca. My apartment complex (Stuy Town) is full of families and dogs; coming home feels like abruptly stepping out of the city and into some unusual country village. Finally, I’ve started using Citibike to get around town. There’s no feeling as freeing as being the fastest thing on the road during rush hour in Manhattan.

My Citibike “hazing” came just a week after I signed up when a mentor of mine who has biked in NYC for decades took me Citibiking from Times Square to Union Square at 5:30 pm. When you’re having a conversation with a fellow biker and there’s no dedicated bike lane, it turns out the move is to weave through traffic erratically, alone half the time, while shouting. It actually worked pretty well, though I decided to buy a helmet after.

The Bad Plus live in NYC

When I moved here in August, I rejoined Jeff Hammerbacher’s group at Mount Sinai to complete my project and dive deeper into computational immunology. Mount Sinai seemed to be the right home base to explore the landscape of biotech in the Northeast. My hope was to validate and narrow down my interests in biomedicine while producing useful science and high-quality software. I love my immunology and genetics work, but know little else in the vast world of biotech and medicine. I wanted to find interesting problems to hit the ground running in grad school a year later, while also shadowing doctors at Mount Sinai and leveraging my proximity to the Boston’s biomedical ecosystem to understand what it is like to bring new therapies to market.

Unfortunately, Jeff’s NYC group just shut down due to some funding issues — a sad reality of academia. That said, it was a positive experience, affording me the opportunity to travel to several conferences and polish off my project (more details above, and here’s the preprint again). In a couple weeks, I’m starting a PM job at Butterfly Networks, a startup that has transformed ultrasound into a ridiculously affordable probe that looks small enough to fit in your pocket. I’ll be working on part of their go-to-market strategy — a fun way to see the inside of the medical devices world while building my PM and customer development skills before I return to research.

This year also comes right as I’ve turned 20 and am starting my life out of college, so it’s time to set and stick to good habits for the decade ahead. I’d like to continue to live intentionally, even when outside the structured cocoon of college. My “Mastermind” accountability group — the subject of a lot of airtime in the last few updates — lives on, though we’re now scattered across multiple continents, meaning updates comes less frequently and are more incisive. I’m hoping to piggyback off the structure of a startup gig to re- emphasize my reading, music, and fitness routines.

I’ve deeply enjoyed meeting new folks in the city — especially musicians and scientists. Still working on building a community and feeling at home here. Please don’t hesitate to suggest folks to meet in NYC or elsewhere in the Northeast, if anyone comes to mind.

2016-7: years of improvised travel on the cheap

Ready for some comparisons of budget bus services at unnecessary levels of detail?

The Puerto Rican rainforest

Here are some recent travel highlightswith photos here (don’t miss Iceland on page 2!):

Officiating my brother's wedding

To top it off: the day after graduation, I hopped on a one-way flight to Europe with only one week planned out and no return date in mind.

Having never traveled solo for a month and a half before, I expected I’d either go crazy or love it. After the fact, I’m still not sure which it ended up being. What’s clear is that the summer of aggressive relaxation was a much- needed break after four long years.

Planning two days ahead at a time was thrilling. I lived with some Senegalese musicians outside Paris, who fed me delicious traditional food and showed me their unique take on traditional Senegalese folk music. Then Airbnb led me to a 3’ x 3’ attic room (that includes the space taken up by a shower) in the heart of Paris’ Latin Quarter. My buddies at Princeton connected me with their European friends and family, and I also reconnected with some former teachers who had moved abroad.

I did my best to breathe in the local culture, and especially focused on rekindling my French. The daily routine: hopping from cafe to bar to cafe, reading and talking to folks around. I can now present an authoritative ranking of schnitzel in Berlin, and I tried all the dark beer I could find in Prague and Amsterdam. Most importantly, I finally understand why Icelandic music has such an ethereal sound. It’s the only music appropriate for driving at the strict speed limit of 90 km/h along deserted two-lane roads through desolation.

Waterfalls in Iceland

Side Projects and Hobbies

Building sustainable organizations on campus: TigerApps, ReCal, and TigerTrek

A few years ago, two friends and I created ReCal, a course scheduler app that went viral on campus. At first, we ran the app off our own credit cards because the undergraduate student body’s hands were tied. They wanted to preserve their close relationships with administrators, who felt frustrated that students flocked to what appeared to be a competitor to infrastructure the university had just poured millions into.

Paying for ReCal ourselves quickly proved unsustainable. Meanwhile, the app had grown to 5,000 users (now we’re north of 6.5k). The engineering school had even started officially recommending ReCal to faculty advising freshmen. My solution: join the undergraduate student government and become an advocate for student developers on campus.

Together with an excellent team of partners in the undergraduate government (USG) and in the computer science department, we built a new TigerApps organization from the ground up in 2016+7. Our USG Labs program provides:

  1. Launch assistance: Plenty of student software projects created for class projects never see the light of day, because launching a project on campus is arduous. We remove the roadblocks in students’ way, providing publicity, data access and APIs, and access to the right administrators.

  2. Funding: we identified a critical funding gap and created a program to provide starter funding to help launch apps for students, by students. Unlike other funders on campus, with whom the conversation inevitably turns to who owns the IP, this money has no strings attached and is solely meant to help students build freely for their peers.

  3. Sustainability: we ended the trend of unreliable student software disappearing as soon as its creators graduated. Instead, successful Labs projects “graduate” into full maintenance by TigerApps. Our containerized infrastructure makes reliability with a rotating cast of maintainers possible.

I quickly learned my strengths and weaknesses as a manager. Recruiting proved to be the biggest challenge — motivating students to work for free calls for some creativity. We found that granting members full ownership over their projects encouraged participation, though our limited budget couldn’t support paying our developers.

Then we ran a PR campaign to nix the image that USG TigerApps was full of deadbeat apps. Our sponsored hackathon prize was calculated to blow the other awards out of the water. This let us immediately incubate some promising apps: an innovative way to find clubs to join on campus, a search tool for study rooms based on the number of devices connected to WiFi in those rooms, a clever scanner app to make consumption of the event posters littered around campus easier, and a student directory browser embedded into the Gmail compose window. Training and positioning the new leader of TigerApps for success also took careful thought and made for a good learning experience for me.

TigerApps Hackathon awards

Really, the TigerApps work boils down to the challenge of how to seed and incite a community. We wanted to unlock the latent student developer community and make prospective CS students come to Princeton because they see it’s a place where they will be supported in their creative endeavors. I’m glad to have gotten a hands-on education in building a community and organization in a low-stakes arena (10 engineers and 5,000 users) forgiving of mistakes.

At the same time, I put time into building another sustainable organization. From the sidelines, I’ve continued working on TigerTrek, Princeton’s annual student trip to Silicon Valley for a week of off-the-record Q&A with founders, VCs, and executives. When I co-led the trip in 2014, I saw TigerTrek as a way for students to discover alternative career paths — an antidote to Princeton’s tradition of sending 50% of grads into finance and consulting. In six years, the trip has transformed into a bustling community, thanks in large part to putting everyone under the same roof for a week. I count fellow Trekkers among my closest friends. So a big priority for me as I departed Princeton was to organize a TigerTrek board of directors with other former trip leaders, in an effort to preserve institutional memory and help advise and grow the TigerTrek program into the future. So far, so good.

Preventing my music from falling into neglect

I played piano in two groups coached by renowned saxophonist Rudresh Mahanthappa. Wednesday evening rehearsals and jams became the highlight of the week. Here’s a clip from one of our concerts:

Wrap Up and Media Recommendations

Thank you for your support and guidance, and for being a part of this thrilling chapter of my life. Hope to keep in touch, and please don’t hesitate to let me know how I can be of help.

— Maxim

P.S. As always, I’ll close with pointers to recent readings and media I loved.


Playlist link:

Highlights to look for: J Dilla; A Tribe Called Quest; Return To Forever; Slum Village; The Bad Plus; Robert Glasper Experiment; Julio Jaramillo. (My convoluted path through the classical and jazz worlds have now brought me into Latin music and hip hop and old school rap. You can trace many artists on this playlist right back to the jazz musicians they were listening to.)




Two pages of photos at

NYC's East River