About

Arpit Gupta is a Robotics Engineer II at 6 River Systems, specializing in perception and software engineering. He has a Master of Science in Mechatronics, Robotics, and Automation Engineering from Worcester Polytechnic Institute and a Bachelor of Engineering in Mechanical Engineering from Madhav Institute of Technology and Science. Arpit also holds a Nanodegree® in Sensor Fusion from Udacity. In his current role at 6 River Systems, Arpit is responsible for building and enhancing the perception pipeline and tools in ROS, C++, and Python for the mobile robot platform. He also works on MLOps and data pipelines for object detection and recognition tasks. Arpit has developed an image annotation tool using Plotly Dash and deployed it using Google Cloud Run. The annotations are pushed to the BigQuery database to be consumed later in the model training stage. He has optimized the extrinsic camera calibration algorithm using custom-written point cloud processing code and developed a Jenkins (CI/CD) pipeline to run the scheduled job in the cloud. Arpit has also developed an event-driven sensor's data buffering and recording tool called snapshot recorder in C++ and ROS. He has worked on sensor characterizations, geometry, and 3D transformations and utilized the DBSCAN clustering algorithm from scikit-learn for lidar range data clustering and artifact matching. Arpit has also developed a blocked camera detection feature using OpenCV and ROS. He has utilized SOLID design principles for test-driven development and wrote C++ unit tests using the gtest library. Arpit has previously worked as a Deep Learning and Vision Intern at Panasonic Automotive, where he worked with the Vision and Sensing team involved in the development of ADAS applications. He successfully estimated the absolute scale of Monocular SLAM using RANSAC for ground plane estimation and implemented a photogrammetry module to measure objects and distances up to 3 cm accuracy by fusing object detection and scale corrected point-cloud. Arpit's tech stack includes object detection, OpenCV, Python, Scikit, CI/CD, test, Jenkins, mobile, deep learning, and TensorFlow. He has 2.94 years of relevant experience and has worked with sensors such as Intel Realsense D435, SICK TIM LiDAR, IMU, ZED Camera, LeddarTech LiDAR, and industrial RGB camera sensors.