ML Research Engineer

About Us

At Prescience, we are building a new healthcare system run by a medical superintelligence.

The bitter lesson has shown us that general methods leveraging computation beat approaches built on hand-crafted human knowledge, every time. Healthcare has ignored this lesson more thoroughly than almost any other industry. Clinical guidelines, prior auth rules, actuarial tables, and care pathways are exactly the kind of human-engineered priors that get steamrolled once you have enough data and enough compute.

Is embedding LLMs into prior authorization software the best use of soon-to-be general intelligence run on a cluster in Abilene? We don't believe so. We sit at the intersection of the three things this moment requires: the technical depth to build frontier AI, the clinical fluency to actually enter the room, and the ambition to own and direct $6T of healthcare spending to ensure Americans don't die.

Prescience is a seed-stage company, most recently valued at $150M post-money. It was founded by Rishab and Aditya Jain, two brothers who have been doing AI research together since 2014. Aditya earned his M.D. from Harvard Medical School and forwent a career in medicine to build Prescience. Rishab was named one of TIME’s 25 Most Influential Teens at 13, took first place in Bioengineering at ISEF, and successfully exited two past businesses before leaving Harvard to co-found Prescience.

Why Prescience?

Prescience is a generational opportunity to rebuild American healthcare from first principles. We have rapid organic adoption from employers, patients, and providers. As AI capabilities improve and the network effects compound across our volume, our business becomes even more indomitable.

In a world with AGI, only three sectors of the economy matter: compute, entertainment, and healthcare. Prescience is building the generational healthcare company, and has massive upside for early believers.

Everyone on the team has engineering backgrounds, and there is no PM layer.

The Role

Our number one priority is scaling to market demand. We are looking for individuals who are high horsepower, high throughput, and extreme resourcefulness to help us increase capacity and grow. We move fast and need to move faster.

The models will do most of the work; your job is to build everything that lets them: the data, harnesses and tools, and infra. This is a founding role with an unprecedented level of ownership and agency. You will pick problems, own them end to end, and ship to production.

We are looking for high agency, high slope individuals. Prior research experience (evidenced by publications or technical blogs) and/or MLE roles are preferred. Here is a preview into the work you will be doing:

• Build data pipelines and find new ways to tokenize and compress medical data;
• Design and execute pretraining runs on longitudinal patient data (Time Machine);
• Train our reinforcement learning system, Crystal;
• Own evals end to end;
• Ship models into prod with real patients. Your research will directly impact P&L;
• Set the research direction. You have an unlimited token budget, and a lot of compute available.

Compensation and Benefits

• A generous cash base.
• Founding-team equity.
• The best health insurance you've ever had.
• Uber vouchers, DoorDash, Equinox.