Third Horizon (Chicago, IL) is hiring an AI Software Engineer to build SLM/LLM interfaces and machine-learning (ML) models that let internal teams and clients query, analyze, and operationalize health care price transparency data at scale.
Requirements
- Proficiency in TypeScript/JavaScript (Node.js, API services, integrations) and Python for SLM/LLM pipelines; SQL/BigQuery familiarity a plus.
- Hands-on with SLM/LLM app frameworks/APIs (OpenAI/Vertex) and open-source models (Hugging Face); prompt engineering, RAG, evaluation.
- Solid understanding of ML/data engineering: data modeling, partitioning/clustering, orchestration, CI/CD, testing.
- Experience on GCP (BigQuery, Cloud Storage, Cloud Run/Functions, IAM, Secret Manager); containerization (Docker).
- Comfortable reading and writing API specs (OpenAPI) and implementing secure, read-only integrations.
- Vector search exposure (pgvector, Pinecone, or BigQuery vector functions) and embedding management.
- Testing frameworks (pytest, jest) and data quality checks (Great Expectations or custom assertions).
Responsibilities
- Build chat/ agent flows answer questions, using retrieval patterns (embeddings/vector search, SQL templates, APIs) in collaboration with analysts and data engineers.
- Design prompt chains, tool use, guardrails, and offline evaluation (gold sets, accuracy/latency/cost metrics).
- Integrate with internal systems: GitHub (read-only), Confluence (docs), and optional BigQuery read-only Actions.
- Develop modular services and APIs using TypeScript/Node.js and Python, integrating with SQL (BigQuery Standard SQL) datasets prepared by analysts and data engineers.
- Package deliverables as reusable libraries or APIs (Cloud Run/Functions) with CI/CD and tests.
- Contribute to Dataform (JS) data processing pipelines in collaboration with data engineers, with attention to performance (partitioning, clustering), correctness, and maintainability.
- Implement QC/metadata reports: stage/shard row balance, directory match rates, code‑set coverage, rate adjustments, outlier bands.
Other
- This position will require in-person work at our HQ in Chicago.
- Bachelor’s or Master’s degree in Computer Science, Data Science, Health Informatics, or related field
- 2+ years building data or ML-backed applications in production (internships/co‑ops count if substantial).
- Excellent communication; ability to translate complex ideas for non‑technical audiences.
- Strong problem-solving abilities and analytical mindset
