Senior Autonomy Controls Engineer – Learning-Based Control

Teleo is a robotics startup disrupting a trillion-dollar industry. Teleo converts construction heavy equipment, like loaders, dozers, excavators, trucks, etc. into autonomous robots. This technology allows a single operator to efficiently control multiple machines simultaneously, delivering substantial benefits to our customers while significantly enhancing operator safety and comfort. Teleo is founded by Vinay Shet and Rom Clément, experienced technology executives who led the development of Lyft’s Self Driving Car and Google Street View. Teleo is backed by YCombinator, Up Partners, F-Prime Capital, and a host of industry luminaries. Teleo’s product is already deployed on several continents and generating revenue.  Teleo is poised for rapid growth. This presents a unique opportunity to be part of a team that is creating a product with a profound impact on our customers, working on cutting-edge 100,000-pound autonomous robots, engineering intricate systems at the intersection of hardware, software, and AI, and joining the early stages of an exciting startup journey. About the Role Own the transition from manually tuned MPC-based vehicle control to learning-driven control policies that adapt across vehicles with minimal human intervention, while maintaining safety and interpretability. Core Responsibilities

Design and implement learning-based control approaches (imitation learning, reinforcement learning, hybrid MPC + learning)

Reduce dependence on hand-tuned control parameters through data-driven methods

Integrate learned controllers into the existing vehicle control stack safely and incrementally

Define interfaces between classical control (MPC, PID, state estimation) and learning-based components

Work closely with the Principal Controls Engineer to translate classical control insights into learning-friendly formulations

Establish validation criteria for learned control policies before real-vehicle deployment

Required Qualifications

Strong software engineering skills in C, C++, or Python (production-quality code)

Deep understanding of modern robotics control systems

Experience with learning-based control or policy optimization for real-world systems

Comfort working close to hardware and real-time constraints

Preferred Qualification

Reinforcement learning or imitation learning for control

Model-based RL, residual learning, or hybrid MPC architectures

Control under uncertainty and partial observability

Debugging and validating control systems on physical platforms

Bonus Points

Experience deploying learned controllers on vehicles or mobile robots

Familiarity with safety-constrained learning methods

Background spanning both classical and modern control theory