Software Engineer Intern / Fellowship (AI / Backend)

PriVii Care is hiring a part-time AI / Backend Engineer Fellow to build and maintain the AI pipeline that turns patient documents, labs, and free text into trustworthy clinical information. The role focuses on extraction, pre-visit conversations, model evaluation, regression testing, and documentation in a healthcare setting.

Responsibilities:

• Own the extraction pipeline that turns patient documents, labs and free text into a defined clinical schema, with the patient's exact wording preserved, every field citing its source, and conflicts marked rather than silently resolved
• Own the pre-visit conversation: question selection, the stopping rule, the question budget, and the handling of what the system must not answer on its own
• Build and maintain the evaluation set and regression tests that tell us whether a prompt or model change made the output better or worse
• Write the specification pages for your own work, so the next engineer can change it without asking you

Requirements:

• Currently pursuing/holding a bachelor's or master's degree in Computer Science, Computer Engineering, or a related field
• Authorized to work in the United States
• Proficiency in Python, and comfortable working in a service codebase rather than notebooks
• Experience with a Python web framework (FastAPI, Flask, or Django) and the basics of designing a REST endpoint: request validation, status codes, and error responses
• Experience with an LLM API, including structured or schema-constrained output. Personal projects count
• Experience with relational databases and SQL, including designing a table and its relationships rather than only querying one
• Automated testing in Python with pytest or equivalent. You own the test suite for your own pipeline, so this is required
• Experience with Git and pull-request workflows
• Ability to commit to a realistic, self-set scope every two weeks and to communicate early when that changes
• Applicants must be legally authorized to work in the United States
• FastAPI and Pydantic specifically, since that's our stack
• RAG, embeddings, or vector search in something you actually shipped
• Systematic evaluation of model output (rubrics, labeled sets, automated scoring) rather than reading samples
• Docker, and deploying a service somewhere other than your laptop
• Experience with clinical or regulated data, or with FHIR, HL7, SNOMED, or LOINC
• Coursework or research in NLP, information extraction, or health informatics