Skills
- Programming Languages - C++, Python
- Software Tools - Git, Docker, Linux OS, GPU Cluster (Lambda - AMD)
- Robotics - ROS Melodic/Kinetic, ROS2 Foxy, Gazebo 9, OpenCV, OpenAI Gym
- Machine Learning - PyTorch, TensorFlow, Keras, Scikit-learn, Pandas, Numpy, Matplotlib, Scipy, RL Stable Baselines
- Hardware - RPLidar, AWS DeepLens, AWS DeepRacer, Turtlebot, Razor IMU
Work experience
Graduate Research Assistant (Jan 2020 - Present)
- Texas A&M University
- Experience :
- Currently working on learning-based autonomous navigation algorithms for unknown environments using a combination of optimal control and reinforcement learning for SLAM, path planning & tracking via onboard LiDAR scan & stereo vision via ROS
- Developed reinforcement learning algorithms for autonomous driving algorithms for 1/18th scale robotic AWS DeepRacer vehicle using Amazon Web Services like Sagemaker/RoboMaker
- Evaluated Proximal Policy Optimization(PPO) for lane keeping & path following robotic vehicle through monocular image data
- Designed PID and LQR waypoint tracking control for 1/18th scale robotic vehicle with A* path planning via ROS Melodic
- Involved in building, training and deploying machine learning and reinforcement learning models for various complex robotic environments in OpenAI Gym using Mujoco or PyBullet Physics Engine.
Advisor: Dr. Dileep Kalathil
Application Engineer (Aug 2017 - May 2019)
- Johnson Controls India Engineering, Mumbai, India
- Experience :
- Led a Project Team for executing Building Automation System (BAS) Projects in North America while ensuring smooth customer communication and delivering quality standards beyond expectations.
- Integrated 3rd party BAS networks with JCI network for Brownfield projects along with configuring an overall BACnet network, programming the control logic, designing the panel layout & controller termination wiring diagrams.
- Implemented energy efficient predictive-maintenance strategies with data-driven analysis & designed user-friendly dashboards