Trajectory
Member of Technical Staff - Research
San Francisco · Staff+
Sponsorship not specifiedDetected 57 days ago
Machine LearningPyTorchLLMsResearchExperimental Design
About the role
- AI is the most capable software ever built, and the least able to learn.
- Every valuable correction and edit that happens in a product evaporates at the next session.
- A few teams have closed this gap by hand-coupling their models to their products: Composer, Claude Code, Windsurf SWE-1.
Responsibilities
- You will own end‑to‑end experiments across data, training, and evaluation: shaping telemetry into learnable signals, training and serving custom LLMs and agents, and designing novel algorithms.
Requirements
- Experience training or serving large language models
- Proficiency in PyTorch, JAX, or similar ML frameworks
Nice to have
- Background in pre- or post-training RL for LLMs
- Experience with high-performance computing or large-scale clusters
- Contributions to open-source ML research or infrastructure
- Demonstrated technical creativity (research, OSS, side projects)
- PhD (last year of PhD), working on AI / RL open source projects
Company info
- our platform unlocks the signal already sitting in product use, so companies can continuously post-train large-scale agentic models that outperform the frontier.
- Working with leading AI companies, including Decagon, Clay, Harvey, Mercor, and Rogo.
- Founded by researchers from Deepmind, OpenAI, Meta, Apple, Amazon, and Scale. $15M led by Conviction, with Fei-Fei Li, Jeff Dean, and founders of Notion, Braintrust, Modal, Hugging Face, and Dropbox.
- As a Member of Technical Staff (Research) at Trajectory, you will design and build the post‑training stack that lets our customers' models continually learn from real production workflows.
Apply directly at Trajectory →Create a free account for alerts like thisView Trajectory immigration profile
This listing is sourced directly from Trajectory's careers page and normalized into a canonical job model.