cv

General Information

Full Name Sanket Rajendra Shah
Date of Birth 8th December 1994
Languages English, German, Hindi

Skill Inventory

  • Deployment
    • TensorRT engine and plugin development
    • CUDA and C++ integration on NVIDIA edge devices.
    • Orin Development with Driveworks, DRIVE OS.
    • Docker, CI/CD pipelines and tests.
  • Development
    • C++, CUDA, OpenCV, LibTorch.
    • Pytorch, Python, Numpy.
    • ROS
  • Conceptual
    • Deep Learning - Lidar,Camera and Radar based 3D Detection, Tracking, Velocity Estimation, Lane Segmentaiton.
    • Computer Vision - 3D computer vision, Structure from Motion, Optical Flow, Point Cloud Processing
    • Robotics - Sensor Fusion, SLAM, Motion Planning, Control Systems,
    • Sensor Setup - Lidar, Camera and Radar Calibration and Synchronization

Experience

  • 2022 - Present
    Perception Engineer
    MotorAI GmbH, Berlin, Germany
    • Deep Learning based Camera-Lidar-Radar Fusion Model.
      • Combines multi-sensor data for 3D Detection and Map Segmentation.
      • From research to training to deployment on vehicle.
      • TensorRT Plugins for Custom CUDA operations.
      • Used C++, CUDA, Pytorch, TensorRT.
      • Presented to BMW Motors.
    • Monocular 3D Detection
      • Performs real-time 3D Detection from monocular images.
      • Trained on public dataset, deployed on vehicle.
      • Converted torch model to TensorRT engine for deployment.
      • Achieved 3x inference boost by quantization.
      • Used C++, CUDA, Pytorch, TensorRT.
    • Traffic Sign Detection
      • 2D German Traffic Sign Detection at 100FPS.
      • Trained on custom dataset and deployed on Vehicle.
      • TensorRT engine generation and deployment in C++.
      • Used C++, CUDA, Pytorch, TensorRT.
  • 2021 - 2022
    Perception Researcher
    Artisense GmbH, Munich, Germany
    • 3D recontruction on Monocular Images.
    • Worked on inhouse research paper implementation.
    • Trained and implemented the model for in-house 4Seasons Dataset.
    • Helped in researching on model performance bottlenecks.
    • Used Pytorch.
  • 2020-2021
    Data Science Werkstudent
    Dinext Pi-Sec GmbH, Saarbrucken, Germany
    • Worked with building PowerBI Dashboards for cybersecutiry.
    • Built over 50 dashboards from scratch for more than 20 German customers.
    • Gave training on Machine Learning using KQL to several German clients.
  • 2019-2020
    Data Science Werkstudent
    Sepago GmbH, Koln, Germany
    • Machine Learning powered PowerBI dashboards for security analytics.
    • Helped in solving real-world on premise cyberattack through machine learning.
  • 2018-2019
    Data Scientist
    Netmonastery Pvt. Ltd., Mumbai, India
    • Developed Machine Learning Models for user behaviour analytics.
    • Worked on software stack to improve data ingestion throughput.
    • Used Python, Pytorch, RabbitMQ, Docker.
  • 2017-2018
    Data Science Intern
    Carnot Pvt Ltd.
    • Applying IoT powered Analytics on tractors to make farming efficient.
    • Worked on wheel slipping prediction ML model for tractors.
    • Used Python, Scikit Learn, Numpy.

Education

  • 2019 - 2022
    Masters in Computer Science
    University of Bonn, Germany
    • Computer Vision, Cognitive Robotics
    • Deep Learning, Refinforcement Learning
    • Humanoid Robotics, Techniques for Self-Driving Cars
    • Masters Thesis, Self-Supervised Learning Based Multi-Object Tracking.
  • 2014 - 2018
    Bachlors in Computer Engineering
    University of Mumbai, India
    • Algorithms, Linear Algebra, Calculus
    • Computer Networks, Operating Systems, Database
    • Machine Learning , AI