Experience

Fusion of ML Engineering & Business Development for SAI Group's portfolio companies

Currently building the future of work (and other buzzwords) at JazzX

Taught lectures and guided teams of masters & PhD students to make conference-ready research papers on multimodal representation learning, modality alignment, translation, generation, fusion and co-learning.

Also previous Head TA of 17-356 Software Engineering for Startups & 15-213 Introduction to Computer Systems

Collecting novel "interesting" data to train Semantic Segmentation & Depth Models for a Self Driving Tractor

Running an Active Learning Experiment and come up with metrics to measure performance of models and methods of data capture

Constructing novel Extensible Smart Contracts & non-fungible tokens (NFTs) on Solidity, allowing off-chain computation and non-deterministic code to be put into use in blockchains

Contributing to the creation of Boba: a Layer-2 Ethereum Scaling Solution (using Optimistic Rollups) reducing the cost and increasing the efficiency of transactions on Ethereum

Studying and creating Automated Market Makers (AMMs), reducing the impermanent loss by comparing token prices on and off- chain and changing the fees earned by liquidity providers.

Exploring Natural Language Processing (NLP) & Computer Vision research problems of prosody transfer, voice cloning, lip syncing to produce a Generative AI voice dubbing solution

Researching State-of-the-Art Unsupervised Speech Decomposition & Generation algorithms like SpeechSplit and identifying applications of disentangling and transforming multiple speakers' content, timbre, rhythm and pitch

Teaching CMU undergraduate students app-development in Android

Creating course material to introduce students to Firebase, building a chat service, using SDKs such as Google Maps, etc.

Developing in-browser Text Detection to identify and remove sensitive patient information from Ultrasound and CT scans.

Modifying existing EAST (Efficient and Accurate Scene Text) algorithm to create a shallower model to identify personally identifiable information (PII) using tensorflow.js, Tesseract and OpenCV.

Developing a B2B e-commerce portal for eyeglasses

Making assistive devices affordable and accessible to opticians esp. in rural areas

Understanding technology behind Aadhaar – India’s Biometric Identity for 1.2 billion citizens

Developing modules for the Aadhaar API and integrating Aadhaar Authentication into local businesses

Organizing inter-school tech symposium ‘ProjectBeta 2.0’ in August 2018

Heading Design & Start-up Pitch Departments and leading teams in inter-school IT events in Graphic Designing, Webdev & Programming