AI Software Engineer,
Clinical AI

About MedBAR.ai

MedBAR.ai built the flight simulator for clinical reasoning.

Medicine never documented its most important layer. The knowledge is in the textbooks, but the reasoning (what to ask, what to order, what to rule out, when to act) was never captured at the scale needed to learn from it. So it has always passed from expert to student one case at a time: at the bedside during residency, and today, in much the same way, in the training of frontier models. That is the layer we build for.

What began as research at the Technion and Rambam is now a live commercial product with leading medical faculties, a national examination body, and top healthcare entities in the U.S. Physicians and the leading AI systems work real, hard cases end to end (history, labs, imaging, diagnosis, plan), along pathways that branch and change, while our engines score every step of the reasoning: quality, safety, efficiency, and a few dimensions we keep to ourselves. The same platform trains and evaluates students, residents and experts, and evaluates the top AI systems and foundation models before and after they go live inside the U.S. healthcare system.

We are a small team built for talent density: scientists who have shipped before, clinicians who know the decisions made at the bedside, and senior healthcare executives from the U.S. and Israel.

Why join

• The stakes are real. Our scores help decide whether an AI system is ready to see patients, and whether a resident is ready to decide alone. Your code is judged in the field, not in a demo.
• No layers. You work directly with everyone in the company and with our close partners, and you own what you build, in production.
• The hard problems are open. Multi-agent systems reasoning through medicine, measured against physicians. Most of what breaks has not been written up anywhere yet. That is the job.

What you'll do

• Design, build and ship LLM-powered features end to end: backend services, agent logic, and the data behind them.
• Build and harden our multi-agent systems: orchestration, prompts, tool use, retrieval, and the guardrails around them.
• Strengthen evaluation and scoring pipelines so results are accurate, reproducible and trustworthy.
• Work directly with clinicians and the clinical testing team, and turn their feedback into engineering changes.
• Own reliability: testing, performance, monitoring, and production issues traced to root cause.
• Take features from idea through production, and keep owning them once they are live.

Must-haves

• Team player and collaborator. No ego. You enjoy building together and make the people around you better.
• Problem solver who likes the fun of it: the interaction, the debugging, the building.
• Ownership and self-drive. You work independently, find the problems, and see things through.
• 3+ years of hands-on AI software development experience.
• Strong experience with AI technologies, LLMs and multi-agent environments.
• Experience at a digital health company or on a healthcare-related product.
• BSc in Computer Science, Engineering or a related field.
• Comfortable in a fast-moving startup environment, and you enjoy it.

Nice-to-haves

• Python/FastAPI backend and React/TypeScript frontend.
• Built a product on top of LLMs in production: you know where they break and what to test.
• Data depth: comfortable going into the details of the data behind a data-driven product.
• Experience working with a clinical or QA testing team.
• Experience with evaluation, scoring or RL-adjacent pipelines.