Staff AI Engineer (AI)

TL;DR

Staff AI Engineer (AI): Building and scaling LLM-powered products and RAG pipelines for healthcare intelligence with an accent on production-grade reliability and cost-efficiency. Focus on developing agents, tool-use systems, and implementing rigorous evaluation suites to eliminate hallucinations.

Location: Remote (Must be based in the US)

Salary: $210,000 – $280,000+

Company

Healthcare intelligence company providing AI-powered platforms to health plans to maximize financial outcomes and reduce costs.

What you will do

• Design and build LLM applications including RAG pipelines, agents, and multi-step workflows.
• Develop and maintain prompt engineering evaluations, treating evals as code.
• Integrate various foundation models (Claude, GPT, Gemini, Llama, Qwen, Mistral) based on use-case.
• Build production-grade infrastructure for caching, streaming, observability, and guardrails.
• Fine-tune or distill models using LoRA/QLoRA, SFT, and DPO when prompting is insufficient.
• Collaborate with product and design to create AI-specific UX patterns.

Requirements

• 3+ years of software engineering experience.
• At least 1 year of experience shipping LLM-based features in production.
• Proficiency in Python and TypeScript.
• Deep knowledge of the LLM stack: function calling, structured outputs, and vector search.
• Experience building and maintaining evaluation suites for AI models.
• Must be based in the US.

Nice to have

• Experience with agent frameworks and Model Context Protocol (MCP).
• Fine-tuning open-weight models or working with inference servers like vLLM, TGI, or SGLang.
• Background in security, red-teaming, or AI evals.
• Contributions to open-source AI tooling.

Culture & Benefits

• Digital-first environment with remote work allowed across the US.
• Comprehensive Medical, Dental, and Vision insurance offerings.
• 401(k) with company match, FSA, and HSA.
• Unlimited PTO and tuition reimbursement.
• Flexible and trusting work environment with high employee empowerment.