Collinear AI
MTS - Research Scientist Internship
Sunnyvale, California +2 · On-site · Internship · Posted Sep 21, 2026
- Location
- Sunnyvale, California
San Francisco, CA
Bengaluru, India - Experience
- The internship is designed for PhD students eligible to do a 12-week internship; no specific number of years is stated.
- Education
- Bachelor’s, Master’s, or PhD in a technical field, or demonstrated proof of work through significant open-source contributions or industry experience.
About the job
Build realistic agent environments, evaluation and verification systems, and data pipelines, and run model post-training to improve the capability and reliability of frontier AI agents.
What you'll do
- Design and implement long-horizon simulated environments for agents.
- Develop policy-aware verification and evaluations, and analyze model failure modes.
- Run post-training and build data pipelines for open-source models.
What they ask for
- Bachelor’s, Master’s, or PhD in a technical field, or demonstrated proof of work through significant open-source contributions or industry experience.
- Strong software engineering foundation and ability to build robust, scalable infrastructure; comfortable in a Python-friendly, CLI-first environment.
- Principled understanding of foundation models, including their construction, evaluation, and optimization.
- Experience conducting reproducible research or technical experiments with data-driven results.
Nice to have
- Experience with Reinforcement Learning (RLHF/RLAIF), simulation systems, or long-horizon agentic environments.
- Influential ML research contributions or high-impact open-source projects; experience fine-tuning or evaluating large-scale models.
Tools
Python, CLI, CPT, SFT, RL, RLHF/RLAIF, MCP
AI focus
Develop agentic environments, evaluation and safety verification, frontier data pipelines, and post-training for reliable AI agents.
Model-written summary · the original posting is authoritative