Lead VLA AI Engineer for Humanoid Robots

CNTXT AI; Abu Dhabi; On-site; AED 240,000 - 320,000; Full time

Job summary

You will design evaluation protocols, run experiments, and produce technical reports. The role requires strong PyTorch skills, experience with LeRobot/openpi/Isaac-GR00T, and a track record in robotics ML; leadership experience is a plus.

Qualifications

• 5+ years of deep learning experience; 3+ years in robot policies or embodied AI.
• Fine-tuning VLA or imitation policies; operate on real robots.
• Strong knowledge of robot imitation learning and RL for robots.
• Expert PyTorch, multi-GPU with mixed precision and LoRA.
• Familiarity with open-source VLA bases (e.g., LeRobot, openpi, Isaac-GR00T).
• Experience with large multimodal datasets, data pipelines and dataset versioning.
• Knowledge of evaluation design: fixed tests, multiple trials, confidence intervals.
• Experience with fast real-time inference: TensorRT, ONNX, quantization.

Responsibilities

• Lead VLA model selection, fine-tuning, and testing with the team.
• Own annotation of robot data; ensure label quality and policy relevance.
• Define evaluation protocols for simulation and real robots; oversee experiments.
• Write technical reports for each experiment and method.
• Collaborate with teams on data capture and annotation requests.
• Mentor junior engineers as the team grows.

Skills

Deep learning

Robot policies

PyTorch

Imitation learning

Reinforcement learning

LoRA

Multi-GPU training

Data annotation

Model fine-tuning

Communication results

Education

PhD in robotics or ML

Tools

LeRobot

openpi

Isaac-GR00T

TensorRT

ROS 2

MuJoCo

Job description

You will design evaluation protocols, run experiments, and produce technical reports. The role requires strong PyTorch skills, experience with LeRobot/openpi/Isaac-GR00T, and a track record in robotics ML; leadership experience is a plus.