We help agent developers maintain and improve their ai software by identifying problems and fixing them with changes to scaffolding code, prompts, and models.

As a Forward Deployed AI engineer, you’ll work with our most high-value users to ensure their software is properly integrated with the platform, our agent is succeeding at working on their codebase, and identifying important new features and capabilities for research and engineering to develop.

You’ll have significant influence over product direction and the utmost autonomy.

You’ll be working on

• Identifying bottlenecks, iterating on the agent product and platform interface, and automating tools for doing so.
• Onboarding clients, assisting the agent in integrating client software into the platform, scoping out new tasks on which to train our agent.

What we’re looking for

• Curiosity, intelligence, clear thinking and communication
• Strong ability to produce good code fast (half will be ephemeral, the other half handed off to research/eng)
• Substantial practical experience using LLMs to generate production software.

Why join us

• Live on the frontier of agent post-training, across multiple fronts. Prompt optimization, SFT, RL, and code synthesis. Work with a team who’s been working on the topic for years.
• Help companies across AI verticals deliver software they’re excited about and solve p0 accuracy and reliability problems for some of the best customers out there
• Operate with full ownership and autonomy

About the interview

1. Initial screener
2. OSS take home assessment OR OSS code review
3. 45 min coding interview
4. Onsite

Python

• postgres, redis, modal, fastapi
• vllm, trl TS
• Next.js

Very open to changing JS if there's a good reason.