My Work
A collection of my projects and professional work.
Reinforcement Learning Training on Unitree G1
Leading the development of RL policies for the Unitree G1 humanoid robot, focusing on complex tasks like walking, arm manipulation, and gripper control. This project utilizes imitation learning and rigorous Sim2Sim testing to ensure policy robustness.
Sim2Real of RL policy
Explored the NVIDIA robotics tech stack to train a reinforcement learning policy for a JetBot in simulation. Successfully deployed the trained policy to a real-world JetBot, demonstrating effective Sim2Real transfer.
Visual and interactive Service Oriented Robot
The Robot detects an empty shelf and alert staff for restocking. Simultaneously. it guides a customer to the Beverages aisle when asked, " where can i find Beverages?"
Facial Emotion Recognition Using CNN
Developed a method to identify seven human emotions using CNN and DNN models. Employed the Haar Cascade Algorithm for face detection and the Dlib library for facial feature extraction.
Page Replacement Algorithm- Os Virtual Lab
Collaborated on a team to build a GUI application for visualizing various page replacement algorithms. Developed functionalities to simulate and demonstrate algorithm behavior, enhancing understanding of their performance characteristics.
Heart Disease Prediction
Led a project to predict heart disease using Random Forest and Logistic Regression models. Implemented the models using scikit-learn to leverage their robust predictive capabilities.