Member of Technical Staff, Model Training (applied research)

What You'll Do8

• Fine-tune and adapt language, vision-language, and robotics models using supervised fine-tuning, reinforcement learning, imitation learning, and distillation.
• Build specialized models for multimodal understanding, robotic perception and control, and autonomous agents that reason, plan, and use tools.
• Create simulation-based training and evaluation environments to study policy learning, task generalization, and sim-to-real transfer.
• Assemble training datasets, reward signals, and learning curricula from expert demonstrations, human feedback, synthetic data, and environment trajectories.
• Develop models that power micro1's flow ecosystem, improving data generation, annotation, quality assurance, and automated grading.
• Feed expert corrections and performance feedback into iterative training cycles to boost accuracy, robustness, and task success.
• Optimize models for production inference, balancing quality, latency, throughput, and compute efficiency against frontier baselines.
• Partner with research and platform engineering teams to move prototypes into scalable, production-ready deployments.

Requirements7

• Solid foundations in machine learning and deep learning, with hands-on experience training, adapting, and evaluating modern AI models.
• Practical experience in at least one area: large language models, vision-language models, robotics, reinforcement learning, or autonomous agents.
• Strong Python and PyTorch skills, with the ability to build reliable training, evaluation, and inference pipelines.
• A track record of translating application needs into measurable model objectives, benchmarks, and performance gains.
• Solid software engineering fundamentals, including building maintainable systems and moving research prototypes toward production.
• Familiarity with model adaptation techniques such as fine-tuning, distillation, imitation learning, or reinforcement learning.
• Strong experimental judgment: independently spot performance bottlenecks, weigh trade-offs, and iterate toward effective solutions.

Who Should Apply

This role suits engineers and researchers with hands-on experience training and adapting modern AI models in Python and PyTorch. You should have depth in at least one area: LLMs, vision-language models, robotics, RL, or autonomous agents. If you enjoy turning research prototypes into production systems with measurable impact, apply.

Salary Insight