New team • Student talent program

VinRobotics Physical Intelligence Lab — Where Machines Learn to Move

Connecting Physics‑Informed ML to Embodiment — where theory meets practice and equations turn into motion.

What we build every day

Dance with Balance

Humanoids walk, run, and balance with human‑like agility — even on uneven terrain or carrying loads.

Hands of Precision

Arms and grippers manipulate safely and efficiently, performing tasks with finesse and reliability.

From Virtual to Reality

Control strategies are perfected in high‑fidelity simulation before taking the leap to real robots.

Challenge the Humanoids

Robots tackle dynamic, real‑world tasks that push their limits and show the future of human‑like motion.

Choose your track

Pick one focus area — or blend them. You’ll ship real features on real robots, mentored by experts.

Dynamics & Whole‑Body Control

  • Model‑based controllers (MPC, QP‑WBC), identification & calibration
  • Adaptive/robust control, torque‑level interfaces

Perception, SLAM & Mapping

  • LiDAR/Visual SLAM in dynamic & large‑scale spaces
  • State estimation, sensor fusion, calibration

Planning & Autonomy

  • Search / sampling / optimization‑based planning
  • Tracking control and behavior trees for loco‑manipulation

Embodied ML

  • Physics‑informed ML, RL/IL, VLA for perception‑to‑action
  • Sim2real, representation learning, policy optimization

Systems & Tooling

  • Realtime C++/CUDA on embedded hardware
  • Dev infra, benchmarking, safety & testing

Human‑Robot Demos

  • End‑to‑end demos for exhibitions & pilots
  • Visualization, telemetry, fleet tools

News

View all

Events

VinRobotics at AI4VN 2025

Highlights from our showcase on mastering core robotics technologies in Vietnam.

Press

Hands-on with VinRobotics’ Humanoid Robot

Third‑gen prototype field notes and capabilities overview.

Mentors & Senior Collaborators

You’ll be supported by an experienced, hands-on crew from academia and industry.

Tran Vu Trung

Dr. Tran Vu Trung

Lead • Motor Control & System Identification

DrEng, Nagoya Institute of Technology, Japan

What we’re working on

Sensor Fusion & Estimation

Sync & calibrate LiDAR, cameras, IMU, encoders; Kalman filters; pose‑graph optimization; modality reliability detection.

SLAM in the Wild

LiDAR/Visual SLAM for dynamic, large‑scale environments.

Mapping

Efficient map representations for navigation in cluttered spaces.

Planning & Tracking Control

Classical & learning‑based planners; robust tracking on hardware.

Code Performance

Clean C++/CUDA for realtime on embedded targets.

System Integration

Behavior trees to coordinate locomotion, navigation & manipulation.

Projects

Coming soon — we’ll list open‑source repos and demo write‑ups here.

Growth path

  • Intern → Innovator: contribute to ongoing projects, master core control & robotics techniques.
  • Expand Expertise: whole‑body control, motion planning, realtime integration.
  • Lead Projects: design & run your own experiments and applications.
  • Shape the Future: build a portfolio for top research/industry roles.

What you’ll gain

  • Direct mentorship from experts and senior researchers.
  • Hands‑on work: make the robot move — from sim to hardware.
  • Deep dives into WBC, MPC, adaptive control & more.
  • A rock‑solid foundation for humanoid robotics careers.

Who we’re looking for

You might be a fit if you…

  • Understand kinematics, dynamics, control & optimization basics
  • Can code in C++/Python/MATLAB for algorithms
  • Are analytical, detail‑oriented, and collaborative

Bonus points

  • Built optimization‑based controllers
  • Realtime or physical robot experience
  • State estimation, identification, or adaptive control

How to apply

We accept rolling applications for semester projects, internships, and thesis placements.

  1. Prepare a short CV and links (GitHub, portfolio, papers).
  2. Tell us your preferred track and why.
  3. Send to vinhumanoid@gmail.com with subject “PIR Talent Program — Your Name”.

FAQ

Is this only about humanoids?

Humanoids are a key platform, but we also work on mobile manipulation, UAV perception, and autonomy — wherever physics & embodiment meet.

Can I join remotely?

On‑site is preferred to get hands‑on hardware time. Hybrid is possible for strong candidates.

What should I know before applying?

Basic linear algebra, calculus, and programming. Familiarity with robotics stacks (ROS, Pinocchio, CasADi, CUDA) helps.

Who will mentor me?

Core leads plus external mentors (AI, UAV, SLAM). You’ll have a primary mentor and a weekly sync.