You'll own the backend intelligence of Bordrs: the systems that read a person's documents, ground themselves in Canadian immigration law, and prepare a file a licensed consultant can review and sign.
What we're looking for
- You've built and maintained LLM-powered features in production, including retrieval, structured extraction, or evaluation.
- You care about accuracy and know how to improve it through good retrieval, clear citations, and testing.
- You're comfortable working with TypeScript or Python and Postgres. Experience with pgvector is helpful, but not required.
- Experience with OCR, document extraction, or validation is a plus.
- You can take a feature from an early idea through to production and work closely with a small team.
What you'll work on
- Improve how Bordrs assesses immigration pathways, flags possible risks, and identifies missing information.
- Build search and retrieval tools based on Canadian immigration law and government guidance.
- Turn uploaded documents into structured information and clear applicant checklists.
- Generate draft government forms for review by licensed immigration professionals.
- Build tests and evaluations that keep these systems reliable as models and requirements change.