Part of the awesome team behind λ SnapStart

- Developed a deep-learning strategy for analyzing digitized histopathology slides of colorectal carcinoma (colon cancer)

- Developed a deep-learning based survival analysis model to predict the Recurrence Free Survival (RFS) among Colorectal Carcinoma patients, improving the previously achieved concordance index by 15%- Deployed a secure federated learning pipeline with Stanford University and Emory University to collaboratively train deep-learning models without sharing sensitive patient radiology data

- Researched patterns in visual aids used by successful Kickstarter campaigns to understand the cognitive and perceptual processing of visualization which lead to successful business proposals- Developed a machine learning tool to identify, classify and tag images containing graphs with a precision of 94% among images scraped from 1000s of successful and unsuccessful Kickstarter campaigns reducing the manual annotating effort by at least 95%

Worked under Prof. Chiranjeeb Bhattacharya on Covid-19 detection using Chest X-rays in collaboration with Niramai Health Analytics- Developed an end-to-end pipeline using a Multi-Task Learning (MTL) based deep learning model which in addition to screening Covid-19 can detect other common lung ailments among patients, generating comprehensive reports highlighting the affected regions in the lung- Deployed a free WhatsApp based service called SWIFT (Simple Whatsapp based Image interpretation at your FingerTips) to provide a swift diagnosis of potential Covid-19 patients by analyzing Chest X-ray images- So far the service has processed more than 700 patient Chest X-ray images received from more than 200 doctors from different parts of India

- Benchmarked inference speeds for several pre-trained neural networks (deployed on DragonBoard 410c) using Arm-NN, Tensorflow-Lite, and Tensorflow Mobile.- Continued researched and development of detection models for on-edge real-time detection on a low-cost microcontroller (DragonBoard 410c)

A robotics research and commercialization internship in collaboration with Carnegie Mellon University’s Robotics Institute - As a part of the ADAS team, researched and prototyped detection models for on-edge real-time detection on a low-cost microcontroller (DragonBoard 410c) - Engaged in interactive webinars by distinguished professors from CMU, in the field of robotics and computer vision. - Gained experience in Agile/Scrum development methodology with 2-week sprint cycles.

- Deployed the Faster R-CNN algorithm to detect vehicles at particular junctions (from aerial videos), compute their speed and draw inferences from the data. - Deployed a linear regression model to learn traffic patterns across various junctions throughout a city. - Devised a workflow to generate an optimized traffic light pattern using the Max-flow Min-cut derived algorithm applied to the above-mentioned parameters. - Devised a workflow to efficiently guide emergency vehicles through an optimized green corridor in case of emergency.

- Implemented the Mask R-CNN algorithm for real-time object detection and classification. - Implemented the LaneNet algorithm for real-time lane detection. - Employed the checkerboard pose estimation technique to calibrate the stereo-camera parameters to extract depth information from the captured images. - Gained experience in time-sensitive project deliveries and demonstration to prospective customers.

An aerospace industry design simulation competition organized by NASA and Boeing engineers to recreate the experience of working on an aerospace company's proposal team- Worked under the counsel of NASA and Boeing engineers in cross-continental collaborated hypothetical companies to prepare industry-level reports to envision space colonies in accordance to an RFP (Request for Proposal) - Designed 3 different space settlements in and around Mars.- Winners of the Asian Semi-Finals as the company “Dougledyne Astrosystems and Flechtel Contructors”- Winners of the International Finals Round as the company “Vulture Aviation”, 1st position amongst finalists from all continents