Homepage

Projects

Spherical Projection for Point Clouds
Traffic Sign Detection
Skin Segmentation
Wound, Umbilicus and Bandage Detection
Semi-Autonomous Robotic System for Remote Trauma Assessment
Lane Detection and Curve Prediction
Object Collection Robotic Arm
Object Collection Robotic Arm
My Creations
Pedestrian Detection
Playing with Neural Networks
Agile Robotics for Industrial Automation Competition (ARIAC)
Path Planning Algorithms

Experience

Aurora Innovation

Senior Software Engineer
March 2023 – Present  

Aurora Innovation

Software Engineer II - Mapping Perception
August 2021 – March 2023  

Leading the effort on developing and productionizing various map products generated by employing state-of-the-art deep learning models, computer vision and linear algebra.

I am also one of the co-leads in the Aurora Fighting Bias in Tech group, where we find and address any potential areas and blindspots 
that might impact the development of our technology and how it connects to diverse consumers.

Aurora Innovation

Software Engineer I - Mapping Perception
March 2020 – August 2021
Working with state-of-the-art deep learning models to autogenerate mapping products.

BITO ROBOTICS

Software Engineer in Computer Vision
August 2019 – March 2020
We develop robotic systems with the goal to democratize robotics. I work with the computer vision team to develop state of the art solutions for the perception system.

Medical Robotics & Equipment Lab

Research Assistant
May 2018 – July 2019
Research Areas
Skin Segmentation: Carried out real-time skin segmentation using deep neural networks like U-Net, Mask-RCNN, etc. Also worked upon creating a custom dataset for training. The published work can be found here.
Wound, Umbilicus and Bandage Detection: Used Faster-RCNN neural network to detect wounds, umbilicus and bandages on trauma patients using a customized and in house dataset.
Robotic Ultrasound: Developed a semi-autonomous tele-operated robotic system to carry out a remote ultrasound. The idea being a radiologist at the hospital would be able to do a remote ultrasound on a patient en-route to the hospital, saving critical and valuable time. The published work cab be found here and here.
Fly-Cycle: Modernized an existing Flycycle robot for immersive game development and rehabilitation purposes mainly for children with muscle dystrophy.

L&T

Larsen & Toubro – Sargent & Lundy

Research Intern
April 2017– June 2017

• Designed and optimized a hold down clamp and pipe assembly for carrying demineralized water in a mega power plant.
• Carried out optimizations like, fillet radius of clamp, length and material analysis in solidworks and Caesar II.
• Used Ansys, to carry out stress analysis and kept the stresses below the allowable stress limit as per the ASTM standards.
• After 17 test runs, the final design parameters were accepted by the company. The thesis report can be found here.

L&T

Larsen & Toubro – Power

Engineering Intern
January 2017– April 2017

• Underwent a four months on-site training on a (660x2 MW) Coal Based power plant.
• Summarized and developed flow charts for different processes throughout the power plant.
• Understood different managerial responsibilities and prepared a report on the power plant.

L&T

Larsen & Toubro – Chyioda

Internship Trainee
May 2016 – July 2016

• Carried out a six weeks study in the field of plant and piping engineering. Underwent a literature review and learned how industrial designing is done.
• Modeled and performed stress calculations of pipes taking into account the established standards and performing material calculations using thickness of pipe and material verification.
• Prepared a technical report for the internship which can be found here.

Publications and Patents

A Robotic System for Remote Trauma Assessment

U.S. Patent. Pub No.: US 2020/0194117 A1

A Semi-Autonomous Robotic System for Remote Trauma Assessment

International Conference on Bioinformatics and Bioengineering, 2019

Modeling and Simulation of a Differential Drive Mobile Robot

International Journal of Scientific & Engineering Research, Volume 7, Issue 5, July 2016

Awards and Recognition

Book Review Committee for "Responsible Software Engineering" - 2023

Reviewed 2 chapters of "Responsible Software Engineering" by Daniel J. Barrett.

logo-international-go-cart-championship

Won the International Go Karting Competition - 2015

We were a team of twenty students representing our college at International Go-Karting Championship (IGC) organized by LPU. We had to completely model, design and fabricate our own go kart for this competition. I was in-charge of the transmissions and the 3D modelling aspects of the go-kart. Additionally I carried out the front impact stress analysis of the go-kart in Ansys.

Our team, "Team Stallions", won the overall trophy bagging a cash prize of 1 lakh rupees. The competition attracted 20000+ spectators with 60 teams participating across the country.

robo

Won the National Robocon - 2014

Nirma Robocon is a team of enthusiastic roboticists at Nirma University. The team takes part in the International Robocon Competition every year, organized by Asia Pacific Broadcasting Union.

I was in charge of designing and optimizing the design for the manual and the autonomous robot. Depending on various task of the competition, I was able to develop four bar gripper, a pneumatic thrower, a ladder climbing mechanism for the autonomous robot, etc.

My team bagged the title of National Winner at National Robocon Competition 2014 competing with 90 other teams from colleges all over India, being the only team able to complete Sabash (Raising of the flag). The competition theme can be found here here and a documentary covering the full competion can be found here.

Blog Posts

In the News

Mentors Speak !!!

I have advised Anirudh in his medical robotics research at University of Maryland from June 2018-August 2019. Anirudh developed Machine Learning techniques for object detection and semantic segmentation for trauma patients. He also developed and tested robotic control algorithms for robotic ultrasound using a 7 DOF manipulator. He has shown to be a quick learner of both new technology and important clinical facts and is intensely scientifically curious. He is self-motivated and diligent in his work and no task was too difficult for him. He has demonstrated exceptional communication skills and is well liked within the Lab.
Axel Krieger
Assistant Professor, University of Maryland
Anirudh was an intern on a team of roboticists working on an autonomous environmental robot. He demonstrated quick understanding and solid leadership skills. So much, so he became the team lead after only a few weeks. Anirudh was given specs for difficult tasks which included 3D design and fabrication of a robotic arm, robot arm analysis, object detection and camera calibration and ROS development. He applied innovative solutions to each problem and delivered on time. Anirudh is a pleasure to work with.
John Crupi
President, Kick Analytics, LLC