Physician AI Engineer

About Counsel

Counsel’s mission is to multiply the world's healthcare capacity. We believe in a future in which every person has a personalized, AI-enabled doctor always in their pocket. To accomplish this, we’re building the first fully-vertical, AI-native medical group from the ground up. We’ve redesigned the entire clinical stack with custom AI tools that make our in-house physicians 10x faster and smarter. [learn more on our blog]

Since our founding in 2023, we have formed partnerships with leading healthcare organizations and are on track to serve hundreds of thousands of patients on our platform within the next year. We have raised over $35M from a16z, Google Ventures, Pear VC and other top-tier VC firms.

Our team combines deep expertise in medicine (Harvard, Mt. Sinai, Mass General, UCSF) and technology (MIT, Google, DeepMind, NVIDIA, Stripe, Meta) — and we’re looking for exceptional people to join us as we scale.

Join us as we build the future of AI in healthcare!

About the Role

Counsel Health is building an AI-enabled care delivery platform, and we need someone who can sit at the intersection of clinical, technical, and product judgment to own its 0→1 build. This is a builder role for someone who can translate real clinical reasoning directly into shippable agents, safety systems, and evaluations that meet a practicing physician's bar for patient care.

The components of clinical AI most likely to harm or help patients, the diagnostic agents, the safety nets, the evals that gate them, can't be built well by engineers alone, physicians alone, or PMs alone. We're looking for one person who is genuinely all three: someone who can scope a clinical problem, design the eval that proves whether it's solved, and ship the code themselves.

What You’ll Do

Own the full arc of clinical and safety agents: scoping with Product and relevant teams, writing the PRD, designing the architecture and evals, building, and shipping to production.

Stand up a clinical eval framework that becomes the canonical quality bar for at least one agent, that’s built to be testable and productionized so any engineer or researcher can run their work against it and know whether it's safe to ship.

Measurably improve core components of Counsel AI (i. e. history-taking, differential diagnosis, triage, treatment planning), benchmarked against physician performance on representative case sets.

Get fluent in the existing stack quickly by reading the code, running the evals, sitting with the clinical team, and surfacing a prioritized list of where Counsel’s AI and safety systems are weakest.

Ship early, guided work end-to-end from shadowing current workflows, picking a tractable agent or safety component, prototyping it, getting it through review, and shipping it.

What You Bring

• MD, DO, or international equivalent, with residency training or 3+ years of clinical experience completed.
• Substantive, hands-on experience building with LLMs in production, beyond just experimentation.

A track record of real ownership over shipped products in industry or startup settings; health tech experience is strongly preferred.

Board-eligible or board-certified status with meaningful active or recent clinical practice (strongly preferred).

Published research, open-source work, or public writing on clinical AI, LLM evaluations, or related areas (strongly preferred).

Experience embedded inside an engineering organization (strongly preferred).

Who You Are

You think like a physician and build like an engineer, meaning you are equally comfortable reasoning through a differential diagnosis and shipping products that operationalizes it.

You've operated at an early-stage startup, or founded one yourself, and you're energized rather than overwhelmed by ambiguity and a fast-moving roadmap.

You have an experimental mindset: you'd rather prototype, test, and iterate than wait for the perfect plan.

You take patient safety personally and are motivated by getting medical AI right, not just getting it out the door.

You're a natural translator across disciplines such that you can make clinical nuance legible to engineers and technical tradeoffs legible to clinicians.

You're drawn to defining your own roadmap rather than executing someone else's spec.

#LI-PINJOB

Recruitment Fraud Notice. Counsel Health only communicates with candidates using official @counselhealth.com email addresses. We will never ask for payment or financial information during the interview process. If you receive suspicious outreach claiming to be from Counsel Health, please contact us at careers@counselhealth.com.

Counsel Health participates in E-Verify. All newly hired employees are processed through the E-Verify system to confirm employment eligibility.