EN DE
Portrait of Pratik Mahankal

Pratik Mahankal

Back-end Developer | AI | Robotics

Hi 👋, I'm Pratik Mahankal. I am a Software Developer. I have worked on projects involving backend development, computer vision and robotics.

Certificates

E-Yantra Robotics Competition

  • 15 April 2021
  • Eyantra, IIT Bombay

Completed the final stage and ranked among the top 50 teams in the Vargi Bots theme. Built a warehouse simulation using robotic arms, multiple cameras, and a web-based monitoring layer.

International DD - ROBOCON 2022

  • 21 August 2022
  • IIT Delhi

Led the end-to-end software development for the robots, including simulation, communication, locomotion, and actuator control. Also contributed to the project documentation.

International ABU - ROBOCON 2021

  • 18 August 2021
  • IIT Delhi

Improved system reliability by building a complete framework for testing actuators and sensors. In addition to the robot software, contributed to the electrical and mechanical aspects of robot construction.

Fabric Analytics Engineer Associate

  • 20 May 2024
  • Microsoft

Plan, implement, and manage data analytics solutions; prepare and provide data; implement and manage semantic models; and explore and analyze data.

Featured Work

Custom Object Detection using Transfer Learning

We created a GUI that lets users capture images, then performed data augmentation and trained the model using ResNet152. The solution also included a detection system, an inventory management interface, and a 2DoF robotic arm simulation.

TATA Power Robot

Developed a 110 kV switchyard maintenance and monitoring robot for cleaning post insulators, greasing isolators, and performing automated surveillance of the switchyard.

Phys.io

Created an application that recommends and monitors physiotherapy exercises. It uses PoseNet for pose estimation and a custom classifier to verify the required pose, count repetitions, measure accuracy, and identify the next movement.

Automated Warehouse Simulation

Built a Gazebo warehouse simulation in which packages are handled by two robotic arms, two logical cameras, a color camera for reading QR codes, and a variable-speed conveyor belt. A web-based monitoring layer displays information about processed packages.