About this role
To bridge the gap between fragile proof-of-concept demos and robust enterprise usage, Starburst needs its first dedicated agent engineer to scale reliable natural language querying for distributed data. Operating with startup speed inside an established company, the product team designs agents that translate user questions into accurate answers grounded in actual enterprise data. The incoming engineer will treat AI agents as production software systems, handling messy real-world conditions like undocumented schemas and diverse access patterns while shipping weekly.
Candidates must bring operational lessons from shipping production LLM systems at an AI-focused product team or startup, bypassing months of trial and error. The ideal person writes code daily, takes full ownership from design to deployment, and establishes high standards for agent reliability, error recovery, and tool orchestration.
What you will do
- Design and build agent architectures covering tool orchestration, planning, memory, multi-step reasoning, and error recovery.
- Ship production-quality agent features with clear ownership from initial design through final deployment.
- Close the gap between demos and production by handling partial schemas, inconsistent metadata, mixed data formats, and unreliable source systems.
- Introduce proven patterns for agent reliability such as structured error handling, fallback chains, observability, and guardrails.
- Build evaluation suites that catch hallucinations, regressions, and grounding gaps before features reach users.
- Collaborate with the AI Research Engineer to integrate grounding and retrieval into agent workflows.
- Drive technical standards for agent development through active coding rather than documentation.
- Turn architectural improvements into measurable gains in task completion rate, accuracy, and time-to-value.
What you need
- Have at least 5 years of experience building AI and machine learning systems.
- Bring a minimum of 1 year of experience shipping agentic or LLM-powered systems in production.
- Possess direct background at an AI startup or AI-focused product team rather than consulting, integration, or agency work.
- Demonstrate hands-on experience with agent frameworks including LangChain, LangGraph, CrewAI, Semantic Kernel, or custom solutions.
- Show production RAG pipeline experience involving chunking strategies, vector databases, embedding models, and reranking.
- Apply strong software engineering fundamentals such as testing, observability, CI/CD, and code review discipline.
- Maintain a track record of shipping frequently and iterating rapidly on user feedback.
Nice to have
- Experience with enterprise data systems including SQL engines, data catalogs, and schema metadata
- Experience building evaluation and observability pipelines for agentic systems
- Familiarity with JVM-based systems or polyglot development environments
- Contributions to open-source agent frameworks or tooling
- Experience with function calling, tool use patterns, and context window management
Benefits
- Competitive pay ranging from $180,000 to $225,000 USD
- Equity packages including ISOs
- Comprehensive benefits offering
- Flexible paid time off
Skills and tech stack
Questions candidates ask
What is the travel policy for this role?
This role requires 25% in-person travel for purposes such as new hire onboarding, team and department offsites, customer engagements, and other company events.
What is the salary range for this position?
The pay range for this role is $180,000 to $225,000 USD, which may be inclusive of variable compensation like commission or bonus depending on location and profile.
What kind of team environment will I be working in?
The team operates with startup speed inside an enterprise company, shipping weekly, measuring results, and building the AI layer for products like AIDA.