Computer Vision/AI Engineer

• Oct 2021 - Present

• Dec 2020 - Sep 2021

Thesis Title- Implement and Evaluation of CNNs for railways detection in Camera and LiDAR modalities.1. Conducted Literature Survey and Conceptualized CNN-based Rail Track detection.2. Implemented a Regression CNN for Rail Track detection in Camera Modality.3. Implemented a Segmentation CNN for Rail Track detection in LiDAR Modality.4. Implemented PyTorch-based utility for Distributed training of CNNs (DDP).

• Sep 2020 - Dec 2020

1. Formulated ROS Diagnostic based ‘Health Monitor’ tool for monitoring Data recording process.2. Established a modular Framework for structuring Deep Learning Projects3. Automated the Visualization and Error Checking process for Annotations of Proprietary Rail Detection Dataset.4. Carried out Data Engineering – Image and LiDAR scans extraction, Adapted Image and LiDAR Annotation, Data Cleaning, Interpolating Ground Truth for camera data etc.

Research Intern

• Jan 2020 - Apr 2020

Focus on Computer Vision and Deep Learning-based HD map verification for Autonomous Vehicles 1. Developed Input pipeline of a neural network for HD map deviation detection 2. Contributed to preprocessing of a dataset comprising of data from LiDAR, and camera sensors 3. Developed 3D visualization of the input data pipeline Focus on Computer Vision and Deep Learning-based HD map verification for Autonomous Vehicles 1. Developed Input pipeline of a neural network for HD map deviation detection 2. Contributed to preprocessing of a dataset comprising of data from LiDAR, and camera sensors 3. Developed 3D visualization of the input data pipeline

Reseach Intern

• Jun 2019 - Nov 2019

Created Dense visual representation for Robotic Grasping of Deformed Objects 1. Implemented the (ResNet based) Dense Objects Nets into Custom Software stack 2. Generated the dataset with ROS based camera interface 3. Transformed raw RGDB data by employing 3D reconstruction, Change Detection and Background randomization into training usable data. Created Dense visual representation for Robotic Grasping of Deformed Objects 1. Implemented the (ResNet based) Dense Objects Nets into Custom Software stack 2. Generated the dataset with ROS based camera interface 3. Transformed raw RGDB data by employing 3D reconstruction, Change Detection and Background randomization into training usable data.

Research Assistant

• Oct 2018 - Mar 2019

1. Carried out Motion Planning for PANDA robot arm with ROS and Moveit! 2. Interfaced planned motion with Gazebo Simulator 3. Created of a warehouse in Gazebo Simulator 4. Interfaced RealSense with Gazebo simulated robot 5. Implemented Pick-Place as requested by the user from ROS 1. Carried out Motion Planning for PANDA robot arm with ROS and Moveit! 2. Interfaced planned motion with Gazebo Simulator 3. Created of a warehouse in Gazebo Simulator 4. Interfaced RealSense with Gazebo simulated robot 5. Implemented Pick-Place as requested by the user from ROS

Design Engineer

• Oct 2014 - Aug 2017

o Developed M-scripts for ADAS/Powertrain feature design o Created Simulink models of ADAS features from Software Requirements. o Successfully automated Simulink processes via MATLAB Scripts o Developed GUI based tool for Automation scripts o Interacted with FCA group to understand requirements and to deliver the output. o Developed M-scripts for ADAS/Powertrain feature design o Created Simulink models of ADAS features from Software Requirements. o Successfully automated Simulink processes via MATLAB Scripts o Developed GUI based tool for Automation scripts o Interacted with FCA group to understand requirements and to deliver the output.