Founding Robotics ML Engineer - VLA & Diffusion Policy

Company: Nimo Technology, Inc. Location: Mountain View, CA (Onsite) Employment Type: Full-time Compensation: $160,000–$250,000 base + equity + benefits (DOE)

About Us

We are building the next generation of intelligent robots powered by foundation models. We're a small, fast-moving team in Mountain View working at the intersection of large-scale robot learning and real-world deployment.

About the Role

We're looking for a Robotics ML Engineer to lead the training and deployment of Vision-Language-Action (VLA) models and diffusion-based policies on our robot platforms. You'll own the full pipeline: data collection → model training → sim/real evaluation → on-robot deployment.

What You'll Do

• Train, fine-tune, and evaluate VLA models (e.g., OpenVLA, π0, RT-2-style architectures) for manipulation tasks
• Develop and improve diffusion policies (Diffusion Policy, ACT, flow-matching variants) for dexterous control
• Build scalable data pipelines for teleoperation, demonstration collection, and dataset curation
• Run large-scale training jobs on multi-GPU clusters; profile and optimize throughput
• Deploy trained policies to real robots and close the sim-to-real loop
• Design evaluation protocols and ablations; iterate quickly based on real-world performance

Required Qualifications

• MS or PhD in Computer Science, Robotics, EE, or related field (PhD or post-Master experience preferred)
• Hands-on experience training VLA models or diffusion policies
• Strong PyTorch skills: custom modules, distributed training, debugging, performance tuning
• Familiarity with imitation learning, behavior cloning, or offline RL
• Experience with robot manipulation data (teleop, demonstrations, multi-camera observations)
• Comfortable working with ROS/ROS2 and integrating learned policies onto real hardware

Nice to Have

• Publications at CoRL, RSS, ICRA, NeurIPS, ICLR, or CVPR in robot learning / embodied AI
• Experience with foundation model fine-tuning (LoRA, QLoRA, full fine-tune at scale)
• Familiarity with simulation environments (Isaac Sim, MuJoCo, RoboCasa, LIBERO, CALVIN)
• Experience with action tokenization, flow matching, or discrete diffusion approaches
• Open-source contributions to lerobot, OpenVLA, or similar projects
• Prior startup experience and willingness to wear multiple hats

Why Join Us

• Work directly on frontier robot learning problems with real hardware in the loop
• Small team, high ownership — your model ships to real robots, not a paper backlog
• Competitive equity in an early-stage company
• Onsite in Mountain View with a tight-knit engineering team