What we're looking for
An NLP Engineer who wants to build LLM systems that run in production, not in a demo. You'll work on RAG pipelines, function-calling agents, and real-time voice, and on the evaluation harnesses and fine-tuning that make them good enough to ship. We train our own small language models, so this is hands-on work on models we own, taken from prototype to live traffic. Solid Python and engineering fundamentals applied to GenAI use cases are essential.
Responsibilities
- Work with LLMs in production, where scalability and latency matter
- Build on open and closed-source LLMs, including function-calling to integrate external tools and APIs
- Build and optimize advanced RAG systems, and low-latency pipelines for real-time voice
- Design and run model evaluation, including LLM-as-a-judge, and continuously improve accuracy
- Build and maintain NLP pipelines that process large-scale text data
- Ship models into the platform alongside product, data science, and engineering
Qualifications
Must have
- Background in Computer Science, Electrical Engineering, Mathematics, or a related field
- 3+ years in NLP development and 3+ years building Python applications in production
- Solid OOP and software engineering fundamentals, and an understanding of the ML project lifecycle
- Proven experience using and evaluating LLMs in production
- Experience running GPU experiments on AWS, GCP, or Azure
- MCP, evaluation harnesses, or LLM observability tooling
- Fluent written and spoken English
Nice to have (we really mean it)