2027 Internship Behavior ML Engineer, Learned Manipulation Policies
Core
Train diffusion flow matching and behavior policies for heavy construction equipment using real fleet data and simulations.
Role type
Intern, Behavior ML Engineer (Learned Manipulation)
Builds
Learned manipulation policies for autonomous excavators and heavy machinery
Domain
Construction robotics / Autonomous systems (via careerplan.io/jobs/96a6423e-7439-4cb4-9109-117b14e47f7c-2027-internship-behavior-ml-engineer-learned-manipulation-policies-at-bedrock-robotics)
Deliverable
production ML models
Required skills
Python, PyTorch, diffusion models, flow matching, imitation learning, behavior cloning, experiment design, failure mode analysis
Preferred skills
VLA models, Reinforcement Learning, sim-to-real gap analysis, large-scale training infrastructure, robotics control
Technologies
PyTorch, diffusion models, flow matching, VAEs
Responsibilities
Train and iterate on diffusion flow matching policies using real and simulated data; build evaluation pipelines for task success and safety; investigate failure modes like distribution shift; collaborate with simulation and eval teams; contribute to shared codebase.
Seniority
Intern
