ML Tech Lead (Motion Planning)

Our team is responsible for the motion brain of the delivery robot: from scene understanding and dynamics prediction to trajectory generation. Our strategy is to combine the development of compact, production-ready transformers with R&D approaches based on an LLM or DriveGPT stack for the next generation of delivery robots.

You will have a rare chance to participate in the creation of a fundamentally new delivery robot module at an early stage. We are looking for a Tech Lead who will take full responsibility for one of the subdomains, build the process, and deliver solutions to the fleet.

What tasks await you

Team management and leading cross-team projects You will need to bring complex tasks to a reasonable logical solution, establish work with adjacent teams, raise the bar for engineering culture, and assist with hiring.

Implementing SOTA approaches You will regularly study the latest approaches, such as Wayformer, PlanTF, MotionLM, DriveGPT4, Diffusion Planner, and adapt them for production-ready solutions.

Improving models with RL You need to develop RL-based pipelines (PPO, GRPO, DPO), build reward models, collect datasets in the simulator, and understand our Sim-to-Real Gap.

Model interpretability You will need to propose ideas for developing model introspection tools to debug complex cases and obtain tools for prompting behavior.

We expect you to have

• Been a Tech Lead or mentor in ML with measurable results and production system launches
• Have extensive experience with DL, transformers, or generative models
• Worked with PyTorch
• Integrated ML solutions into real products

Will be a plus

• Developed and implemented AV-ML solutions, possess knowledge of Imitation and Reinforcement Learning, SLAM, Motion Planning technologies
• Worked in robotics, familiar with concepts of ROS, Jetson, "sensor set"
• Participated in ML or programming competitions: ACM ICPC, Codeforces, Topcoder, Kaggle, etc.
• Can write production-grade code in C++