AI Infrastructure Intern

Job Description

AI Engineering Intern | Heka Intelligence

The Tone:
This is an unpaid internship at Heka Intelligence, located in the San Francisco Bay Area. Heka is a stealth-stage startup building AI infrastructure for clinical referral workflows. This role is critical for a student eager to learn how to construct structured, tested, observable, and safe AI systems that are ready for deployment in real healthcare environments. It offers direct mentorship and a unique perspective on building an early-stage company in a rapidly evolving field.

The TL;DR
• Role: Internship
• Location: In-person, San Francisco Bay Area
• Team: Founding team (Stanford, Harvard, Cornell alumni) offering direct access and mentorship
• Mission: Learn to build applied AI systems that are structured, tested, observable, and safe for clinical referral workflows.
• Tech Stack: Python, TypeScript, Node.js, Pydantic, Zod, TypeBox, AWS Bedrock, Temporal, Postgres, S3, SQS, Lambda, ECS, Docker, GitHub Actions, DSPy, GEPA-style optimization, OpenTelemetry, CloudWatch

What You’ll Actually Do
• Structured Outputs: Implement structured LLM outputs utilizing Pydantic, JSON Schema, and validation tooling.
• Document AI: Develop and optimize extraction and classification pipelines for document AI.
• Prompt Optimization: Construct eval datasets and apply eval-driven prompt optimization techniques like DSPy or GEPA-style methods.
• Workflow Orchestration: Design and integrate workflow orchestration within Temporal for AI systems.
• AI Observability: Establish AI observability features including traces, confidence scores, human corrections, and regression reports.
• Deployment Safety: Build tests and deployment tooling to ensure AI changes can move safely toward production.

The Must-Haves
• Background: Current undergraduate or graduate student in Computer Science, Software Engineering, AI/ML, Data Science, or a related field.
• Experience: Demonstrated ability to build projects outside of academic coursework, combined with comfort working with APIs, databases, and backend systems.
• Skills: Strong programming proficiency in Python, TypeScript, or both; a keen interest in LLMs, AI agents, evals, document processing, or workflow automation; good judgment regarding privacy, reliability, and correctness; and excitement for applying AI to healthcare challenges.
• Bonus: Prior experience with LLM, RAG, agent, or ML projects; familiarity with FastAPI or typed-validation tools; exposure to cloud platforms (AWS, GCP, Azure); experience with Docker or CI/CD; work with OCR, PDFs, or data pipelines; or coursework in Machine Learning, Natural Language Processing, databases, distributed systems, or algorithms.