Sr. Software Engineer, GenAI

Instrumentl is a profitable, hypergrowth, YC-backed SaaS platform building the operating system for grant-funded organizations. The Senior Software Engineer will join the AI Engineering team to own AI features for grant discovery, from data processing to production deployment.

Responsibilities:

• Build tool-using LLM systems that plan, call tools, and run multi-step workflows for tasks like grant discovery, data ingestion, and research assistance
• Build agentic data-processing pipelines that crawl the web, pull and dedupe messy source data at scale, and structure it into clean, queryable databases other teams build on
• Turn prototypes into resilient production services with clear fallback, cost, and latency budgets
• Own RAG end to end: ingestion, chunking and embedding strategy, hybrid retrieval, re-ranking, citations, and grounding
• Build ranking and scoring systems that match grants to nonprofits, universities, and foundations using complex relevance techniques
• Continuously improve recall and precision and keep indices healthy as the dataset grows
• Stand up evaluation and observability so our AI is grounded, safe, and cost-effective, and treat LLM behavior as non-deterministic by design rather than as a regular API
• Partner directly with founders and your pod on undefined, complex problems with real autonomy
• Write clear, maintainable, well-tested code and build reusably so your work expands across teams

Requirements:

• 7+ years of professional software engineering experience, with deep, recent, multi-year Python and strong relational database and schema design skills
• Solid CS fundamentals and a demonstrated track record of owning complex systems end to end, from design through production reliability
• At least 1 year of hands-on experience building with modern LLMs (as an IC)
• Real RAG depth: hybrid search (keyword plus vector), re-ranking or fusion methods, and grounded citations, tuned in production rather than read about
• Hands-on with at least one of LangChain, LangGraph, or LlamaIndex
• Vector databases beyond pgvector (Pinecone, Qdrant, Milvus)
• Built end-to-end agentic data-processing systems (crawling, dedup, structuring) with whole-system ownership
• Evaluation and observability for AI systems: golden datasets, precision vs. recall, LLM-as-judge, and drift monitoring
• Ruby on Rails (our core platform is on Rails), deep SQL, and experience with AWS or GCP, Docker, and CI/CD
• Startup experience and comfort operating in fast, scrappy, low-process environments