AI Engineer (Robot Learning)

Overview

At WIRobotics, we develop a new paradigm of robotic systems by combining deep understanding of physical interaction with AI-driven intelligence. We build large-scale robotic AI models that integrate vision, language, and action, enabling humanoid robots to physically interact with the real world and learn from experience. Our goal is to realize Physical Intelligence—robots with human-level intelligence, manipulation, and adaptive behavior beyond the limits of conventional robotics.Key Responsibilities.

Key Responsibilities

• Apply state-of-the-art robot learning models such as ACT, Diffusion Policy, and VLA (Vision-Language-Action) to real robotic systems and continuously improve their performance. Work on end-to-end systems, including training, deployment, and debugging of models in real-world environments.
  • Apply and train ACT, Diffusion Policy, and Transformer-based robot policy models
  • Implement and operate VLA-based action policies
  • Train control policies using Imitation Learning and Reinforcement Learning
  • Apply algorithms such as DAgger, Behavior Cloning, and Policy Gradient methods
  • Design robot data collection pipelines and perform data curation, cleaning, and augmentation
  • Build robust training and validation pipelines across simulation and real robots

Qualifications

• Strong understanding of machine learning and deep learning models, including Transformers, Diffusion models, ACT, and VLA
• Solid understanding and hands-on experience with Imitation Learning (IL) and Reinforcement Learning (RL) algorithms such as Behavior Cloning, DAgger, and Policy Gradient methods
• Experience in model development and optimization using Python and C++
• Experience deploying trained models to real robot control systems

Preferred Qualifications

• Advanced degree (Master’s or Ph.D.) in Robotics, Machine Learning, Control, or a related field
• Experience with robot learning simulators and benchmarks such as Isaac Sim, MuJoCo, RLBench, and RT-X
• Experience utilizing and tuning VLA action spaces based on multimodal policy architectures

How to Apply

Email: recruit@wirobotics.com Documents: Resume, Portfolio (optional) Deadline: Open until filled