LLM / GenAI Engineer — Seattle, WA

About The Role

The role builds production-grade LLM and generative at systems: RAG architectures, agentic workflows, fine-tuning pipelines, and evaluation frameworks that move models from demo to dependable.

This team sits at the intersection of applied research and platform engineering. You will ship GenAI features that real users depend on, where latency, cost, and output quality are all first-class constraints.

Key Responsibilities

• Design and implement RAG pipelines end-to-end: chunking strategies, embedding selection, hybrid retrieval, reranking, and context assembly

• Build agentic workflows and tool-using LLM applications with LangChain, LlamaIndex, or custom orchestration, including guardrails and failure handling

• Develop systematic evaluation frameworks: golden datasets, LLM-as-judge pipelines, regression test suites, and online A/B measurement of output quality

• Run parameter-efficient fine-tuning (LoRA, QLoRA) and distillation on domain datasets, managing the full experiment tracking and model registry lifecycle

• Optimize inference performance and cost: quantization, batching, caching strategies, and self-hosted vs. API-based model tradeoffs across OpenAI, Anthropic, and open-weight models (Llama, Mistral)

• Integrate vector databases (Pinecone, Weaviate, pgvector) and streaming data pipelines to keep retrieval indices fresh at production scale

• Collaborate with product and backend engineers to ship GenAI features, participating in design reviews and setting engineering standards for LLM reliability

What We Are Looking For

• 3–6 years of software engineering experience, with at least 1–2 years building LLM/GenAI systems in production

• Strong Python skills and hands-on experience with PyTorch; comfort with async programming, REST APIs, and cloud infrastructure (AWS, GCP, or Azure)

• Proven experience building RAG or agentic systems that served real users, not just prototypes

• Working knowledge of embedding models, vector search, and semantic retrieval tradeoffs in production environments

• Experience with LLM fine-tuning (instruction tuning, PEFT/LoRA) or serving optimized open-weight models

• MS in Computer Science, or equivalent practical experience; strong written communication and system design fundamentals

• Bonus: experience with Kubernetes and GPU inference (vLLM, Triton), RLHF or preference optimization, observability tooling for LLM apps (LangSmith, LangFuse), or contributions to open-source LLM projects