Ensure you read the information regarding this opportunity thoroughly before making an application.
Heka is a stealth-stage startup building AI infrastructure for clinical referral workflows. We're early but real. We already have three clinics lined up for pilot deployments. We're hands-on, moving fast, and building the patterns for clinical AI that the rest of the industry hasn't written down yet.
THE TEAM
You'll work directly with our founding team: a small group of Stanford, Harvard, and Cornell alumni with backgrounds spanning AI/ML engineering, quantitative research, and healthcare. Because we're small and early, an intern gets unusually direct access here, real code review, hands-on mentorship, and a front-row view of how an early-stage company actually gets built.
ABOUT THE ROLE
A mentorship-driven internship for a student who wants to learn how applied AI systems are actually built: structured, tested, observable, and safe enough to run in real healthcare settings. You'll work closely with the founding team, with direct mentorship and code review, on the AI systems behind our product. The goal is your learning: you'll leave with hands-on experience and a portfolio most students never get.
WHAT YOU'LL WORK ON
(with mentorship and review)
You'll get hands-on across the AI stack behind our product:
Structured LLM outputs using Pydantic, JSON Schema, and validation tooling
Document AI: extraction and classification pipelines
Eval datasets and eval-driven prompt optimization (DSPy / GEPA-style)
Workflow orchestration in Temporal
AI observability: traces, confidence scores, human corrections, regression reports
Tests and deployment tooling so AI changes can move safely toward production
TECH YOU'LL BE EXPOSED TO
Python · TypeScript / Node.js · Pydantic · Zod / TypeBox · AWS Bedrock · Temporal · Postgres · S3 / SQS / Lambda / ECS · Docker · GitHub Actions · DSPy / GEPA-style optimization · OpenTelemetry / CloudWatch (You won't touch all of these; we'll pick what fits your projects.)
WHAT WE'RE LOOKING FOR
A current undergraduate or graduate student in CS, Software Engineering, AI/ML, Data Science, or a related field, ideally able to take this for academic credit or through a co-op / work-integrated-learning program. You should have:
Strong programming ability in Python, TypeScript, or both
Experience building projects outside of class
Interest in LLMs, AI agents, evals, document processing, or workflow automation
Comfort with APIs, databases, and backend systems
Good judgment around privacy, reliability, and correctness
Excitement about applying AI to healthcare
You do NOT need prior healthcare or full-time engineering experience. xsgimln
NICE TO HAVE
LLM / RAG / agent / ML projects · FastAPI or typed-validation tools · cloud exposure (AWS/GCP/Azure) · Docker / CI/CD · OCR, PDFs, or data pipelines · coursework in ML, NLP, databases, distributed systems, or algorithms.
WHAT YOU'LL GAIN
Direct mentorship from the founding team, with regular code review
Hands-on experience with industry-standard applied AI in a real clinical setting
A concrete portfolio you can talk through in interviews
A letter of recommendation and a reference on successful completion