MTS - Research Scientist Internship
Core
Build high-fidelity simulation environments and evaluation stacks for frontier AI agents to stress-test reliability and safety.
Role type
Research Scientist Intern (Applied Scientist)
Builds
SimLabs (ultra-realistic simulation environments), evaluation judges, post-training data pipelines
Domain
Artificial Intelligence, Machine Learning, Agent Systems
Deliverable
production ML models
Required skills
Python, software engineering, foundation models, reinforcement learning, simulation systems, data pipelines, model evaluation, open-source contributions
Preferred skills
publications at NeurIPS/ICLR/ICML, RLHF/RLAIF experience, long-horizon agentic environments, fine-tuning large-scale models
Technologies
Python, CLI
Responsibilities
Design and implement SimLabs simulation environments; Develop policy-aware judges and evaluations; Execute post-training runs (CPT, SFT, RL); Analyze model failure modes and create data pipelines
Seniority
Intern
