Job description
We are seeking a
Physical AI Engineer
to design, develop, and implement AI-driven control and decision-making systems for humanoid robots and embodied agents.
This role involves integrating
vision-language-action models, reinforcement learning, imitation learning, and real-time robotics systems
to create robots capable of performing complex tasks in dynamic environments.
Key Responsibilities
Integrate
AI models (transformers, diffusion policy networks, LLMs, vision-language models)
with physical humanoid robots.
Design real-time
control frameworks
that enable AI decision-making to translate into smooth, safe, and efficient motor actions.
Develop pipelines to align
simulation-to-reality (sim2real)
and optimise robot learning for real-world deployment.
Apply
multi-modal AI learning
(vision, audio, haptics, proprioception) to enhance robot perception and adaptability.
Collaborate with hardware teams to calibrate robot sensors, optimise energy efficiency, and ensure reliable AI control execution.
Develop
safety protocols
for autonomous decision-making in environments with humans.
Conduct experiments in areas such as
human-robot interaction, autonomous navigation, dexterous manipulation, and multi-agent collaboration .
Research and implement
emerging paradigms in Physical AI , including embodied GPTs, action-conditioned transformers, and world-model-based learning.
Required Qualifications
Bachelor’s or Master’s degree in
Robotics, Mechatronics, Computer Science, AI, or a related field
(Ph.D. preferred for senior positions).
Strong background in
machine learning , particularly reinforcement learning, imitation learning, or embodied AI systems.
Hands-on experience with
robotics simulation environments
(Isaac Sim, MuJoCo, Unity, Gazebo).
Proficiency in
Python
and good knowledge of
C++/ROS2
for robotics integration.
Deep understanding of
robotics control theory, kinematics, and sensor fusion .
Demonstrated work with
humanoid robots, quadrupeds, or robotic arms , preferably on vision-language-action models.
Preferred Skills
Familiarity with
NVIDIA Jetson, CUDA optimisation, and real-time robotics inference .
Experience with
large foundation models
adapted for robotics (OpenAI VLA, Google RT-2, Gr00T).
Knowledge of
safety-critical autonomous systems .
Exposure to
distributed training
of large models on RTX/HPC clusters.
Working knowledge of
teleoperation frameworks
(for collecting demonstration data).
What We Offer
Salary Range:
₹15-20 LPA (Lakhs per annum)
The chance to work on
cutting-edge humanoid robotics with embodied AI .
Access to advanced computing infrastructure ( NVIDIA RTX 6000 GPUs, Jetson Thor platforms, VR mocap systems, and Unitree robots ).
A multi-disciplinary team environment spanning
AI research and robotics engineering.
Competitive remuneration and growth opportunities for leadership in
next-gen Physical AI projects .
Required Skill Profession
Engineers