Our core product is explainable semantic matching between company capabilities and live RFPs. You will work on retrieval, ranking, and evaluation, not generic chat wrappers.
You will partner with engineering on ingestion quality, offline evals, and production monitoring of relevance.
What you'll do
Design and tune retrieval pipelines over procurement documents and customer-uploaded profiles.
Build evaluation sets and regression tests for relevance, citation accuracy, and latency.
Experiment with chunking, metadata filters, and reranking strategies for long solicitations.
Ship improvements to match explanations (cited requirement spans shown to users).
Document model and index changes so the team can reproduce results.
What we're looking for
2+ years applied ML or search engineering in production.
Hands-on experience with embeddings, vector stores, or hybrid retrieval.
Python proficiency and comfort reading research papers and translating ideas into shipped code.
Structured approach to offline metrics before rolling out ranking changes.
Ability to explain trade-offs to non-ML teammates and customers.
Nice to have
Experience with RAG evaluation frameworks (e.g. Ragas-style workflows).
Knowledge of government contracting terminology (NAICS, FAR, evaluation criteria).
Familiarity with prompt structuring for extraction tasks (not open-ended chat).
J-18808-Ljbffr