Robotics Engineer focused on autonomous systems, robot perception, reinforcement learning, and human-robot interaction.

M.S. in Robotics Engineering from the University of Michigan-Dearborn. I build robotics systems using ROS 2, MuJoCo, Python, C++, and simulation-driven development workflows.

• Robotics software development with ROS and ROS 2
• Autonomous navigation and motion control
• Reinforcement learning for robotics
• Robot perception and sensor fusion
• Probabilistic localization and state estimation
• Human-robot interaction and multimodal robot behavior
• Simulation-driven robotics development with MuJoCo and Gazebo

Modular ROS 2 perception pipeline for real-time PointCloud2 processing, clustering, visualization, and lightweight object tracking.

• Built a ROS 2 Humble pipeline for point cloud preprocessing and ground segmentation
• Implemented Euclidean clustering with 3D bounding boxes in RViz2
• Added centroid-based tracking with persistent object IDs
• Included runtime performance logging, YAML configuration, and unified launch support
• Tech stack: ROS 2 Humble, C++, Python, PCL, RViz2, Ubuntu 22.04

Repository: https://github.com/ninad164/ros2-pcl-perception-suite

Master's thesis project focused on multimodal conversational robotics and trust-aware HRI systems.

• Built a ROS-based multimodal companion robot on Clearpath JACKAL
• Integrated Whisper, OAK-D Pro, DeepSpectrumLite, and local LLM inference
• Conducted an IRB-approved HRI study with 60 participants
• Evaluated trust, adaptability, and emotion recognition in human-robot interaction
• Tech stack: ROS, JACKAL, OAK-D Pro, HRI, Python, LLMs

Reinforcement learning project for autonomous robotic navigation in simulation.

• Built a custom MuJoCo navigation environment
• Trained PPO agents for obstacle avoidance and goal-directed navigation
• Compared learned policy performance against a baseline controller
• Evaluated success rate, collision rate, and trajectory efficiency
• Tech stack: MuJoCo, Reinforcement Learning, PPO, Gymnasium, Stable-Baselines3, Python

Repository: https://github.com/ninad164/mujoco-rl-navigation

Extended Kalman Filter based robotic localization project for probabilistic state estimation under noisy motion and sensor conditions.

• Implemented a nonlinear prediction-update localization pipeline
• Performed robot trajectory estimation under uncertainty
• Demonstrated sensor fusion and motion modeling concepts
• Evaluated localization robustness in simulation
• Tech stack: EKF, State Estimation, Sensor Fusion, Robotics, Python, C++

Robotics: ROS, ROS 2, Gazebo, RViz, MuJoCo, MoveIt 2

Perception: OpenCV, Point Cloud Library (PCL)

AI / ML: PyTorch, Reinforcement Learning, Computer Vision, LLMs

Programming: Python, C++, MATLAB

Robotics Concepts: SLAM, Localization, Sensor Fusion, Robot Navigation, EKF

Portfolio: https://my-portfolio-one-eta-83.vercel.app/

LinkedIn: https://www.linkedin.com/in/ninadalurkar

Phone: +1-248-990-5119