Mountain View, California, United States
I believe certain predilections could be strongly driven by pursuits that are rewarding. Though my tryst with computing technology began a decade ago I still stand amazed and inspired on witnessing the ability of machines to mimic Human senses, and its immense analytical capabilities. Therefore, I am a dedicated Machine Learning researcher and developer, with the goal to intelligently model every aspect of human life through AI, and give back to society through valuable & insightful tech. Most of my work spans the domain of Natural Language Processing & Computer Vision and I have successfully applied my scrupulous data analysis skills to multiple Enterprise level tasks. Along with many research projects and publications in the field of ML, I have participated in multiple Hackathons and Appathons, with an effort to bring something new with every stint. Technologies such as Data Analysis, Knowledge Mining, Computational Linguistics and Vision interest me. My projects and exposure to industry applicable tasks, along with Government sponsored fellowships for research, motivate me to strive towards establishing myself as a leading figure in the field of ML. I believe I can get up to speed with any technology in no time, and develop skills as required of me.
Working with the Advertising Technology team to research and develop ML solutions for forecasting and optimization for e-commerce advertising. - Deep Learning and Transfer Learning - Foundation Models for Time-Series and Language - Sponsored Products optimization - Media/Marketing Mix Models
Working with the Advertising Tech Data Science team to research and develop efficient distributed algorithms for data driven ML solutions for advertising products to e-commerce users
• Collaborated with the Government of New York (and NYCHA) to build NLU based automated decision system for policy making • Worked with Prof. Kathleen McKeown (NLP Group) and the School of Social Work (SAFE Lab) to create Multimodal Stance Detection systems for Public Policies • Researched and implemented transformer-based adapter architectures to achieve state-of-the-art results on zero-shot dataset
• Assisting with instruction and logistics for the Machine Learning course by Dr. John Paisley, of the Columbia-edX AI Micromasters Program. • Helped and guided learners to strengthen ML concepts and course related topics including Statistics, Linear Algebra and Probability.
• Graduate Teaching Assistant for for the NLP course by Prof. Daniel Bauer. • Helped, guided and graded learners to strengthen ML/NLP concepts and course related topics.
• Working with the Advertising Tech Data Science team to research and develop efficient distributed algorithms for data driven ML solutions for advertising products to e-commerce users • Introduced and implemented Contrastive Learning based unsupervised techniques and item Negative Sampling to generate user-item vector embeddings, achieving over 95% acc. in item neighborhood taxonomy and over 92% acc. in new-user item recommendation • Designed effective ways to inject demographic information into User Embeddings using vector projections with triplet loss training • Synthesized data into distributed datasets to support and utilize the analytics project lifecycle to drive modeling using Scala, PySpark
• Graduate Teaching Assistant for "Data Analytics in Python" under Dr. Hardeep Johar for a class of 40 MBA students. • Held weekly Office Hours to help and guide learners in terms of programming and course related topics.
Machine Learning for predicting clinical events. Full Stack Software Engineering. • Conceptualized and engineered 50+ Deep Learning models for predicting clinical events including: Sepsis, Mortality & COVID-19 • Attained 95% AUROC for Sepsis detection, 24 hours before onset; and 98% AUROC for in-ICU Mortality within a 48-hour window • Accelerated data preprocessing execution time, attaining 60X speedup leveraging vectorization in Python, NumPy & Pandas