You build with models for a living. Agents, retrieval, tool-use, and the systems that make them reliable.

What you’ll do

Build agentic and LLM-powered features that hold up in production.
Design retrieval, tool-use, and orchestration.
Write evals so quality is a number, not a feeling.
Know when to reach for a model and when plain code wins.

Your stack

Python, PyTorch, LLM APIs, RAG, vector DBs, and eval harnesses. Or your equivalent. We care about the ability, not the exact tools.

The one hard requirement: you are AI-native

You do not just use AI tools, you build with them. You are fluent with agentic coding (Cursor, Claude Code, Copilot) and you ship faster because of it, not despite it. You have taste about when to trust the model and when not to, you write checks or evals for anything that matters, and you turn vague problems into shipped software in days. If AI coding is still a novelty to you rather than how you already work, this is not the role.

You are probably a fit if

You have shipped something real with models, not just prototyped.
You read the failure cases, not just the happy path.
You are suspicious of demos, including your own.

Logistics

Remote, global. Full time. Total compensation is $180k to $400k depending on level and location.

What stays open

How you end up working with us, a project, a role on a team, or something in between, is a conversation. Apply and we will figure out the shape.

More about the ai / ml engineer track →