Senior AI Engineer, Agents_TT

We're looking for a Senior AI Engineer to take production clinical AI agents to the state of the art at a healthtech company building AI Healthcare Employees that work invisibly inside clinical workflows across Latin America. You'll design clinical ontologies, fine-tune small LLMs, and build robust evaluation frameworks that push agent quality well beyond standard metrics. This is a highly technical, hands-on role: you'll train and fine-tune models end-to-end, reason carefully about when to use context engineering versus fine-tuning, and ship directly to production. The role calls for deep Python skills, fluency with RAG and knowledge graphs, and daily use of AI coding tools to move fast. You'll be expected to own the agent stack with extreme ownership and a genuine obsession with getting the details right. This is an individual contributor path for someone who wants to stay hands-on rather than move into management.

Details:
Schedule: Full time
Location: Remote (Latin America)
Start: ASAP
Duration: Long-term
English: Fluent
Type of collaboration: B2B Contract

About the project:
The project is a healthtech company born out of Stanford that builds AI-powered "employees" designed to work invisibly inside clinical workflows across Latin America. The team runs production AI agents that support real clinical work, and the focus now is pushing those agents to the state of the art: better ontologies, better fine-tuned models, and evaluation that catches failure modes standard metrics miss. The culture rewards extreme ownership and deep technical craft over process, and individual contributors are expected to own their piece of the agent stack end to end rather than hand it off. It's a fast-paced, highly technical environment similar to other leading clinical AI companies, where engineers are judged on the quality of what ships to production.

You have:
- Based in Latin America
- Shipped and maintained AI agents or LLM-based systems in production
- Hands-on experience training, fine-tuning, or distilling models end-to-end
- Strong Python programming skills and experience with context engineering (RAG, vector stores, knowledge graphs)
- Nice to have: healthcare or clinical domain experience
- Nice to have: based specifically in São Paulo, Bogotá, Mexico City, Buenos Aires, or Chile
- Nice to have: recent tenure at a fast-paced, early-stage startup

What to do:
- Design clinical ontologies and knowledge graphs that ground agent behavior in real clinical context
- Fine-tune and distill small LLMs for production clinical use cases
- Build and maintain evaluation frameworks that surface agent failure modes beyond standard metrics
- Decide between context engineering (RAG) and fine-tuning based on latency, cost, and quality tradeoffs
- Own the agent stack in production, including diagnosing and fixing issues you ship
- Use AI coding tools daily to move faster and get leverage on your own output