I am a Computer Science graduate student at the University of Southern California and a research assistant at the Information Sciences Institute(ISI). I am currently working with knowledge graphs and natural language processing at ISI. Before joining USC I was Technical lead at inVOID technologies, where I created state of the art facial recognition model and backend systems for vision-based ID Verification. I have experience working with following technologies:
Language: Python, Java, JavaScript, SQL
Data Science: Spark, MapReduce, Pytorch, Tensorflow, AWS, NLP, Machine learning, D3.js.
Web Dev: NodeJS, Flask, Django, PostgreSQL, MongoDB, NoSQL, ReactJS, HTML, CSS, RESTful API
I am currently looking for Co-op opportunities for Spring 2021 Or full time oportunities starting June 2021. I can work full time with CPT/OPT work authorization on my F1 visa
-
Experience
Research Assistant
Information Sciences Institute, USC
Nov 2019 – Present
• Working with knowledge graph division on multiple research projects part time with graduate studies.
• Created novel transformer NLP architecture to combine features from different text sources like webpages and patents.
• Improved prediction capabilities by 14% as compared to previous architecture. And reduced processing time by 400%.
• Created Web archive index and spark ETL pipeline to process terabyte scale(3.5TB) web text.
• Gaining Knowledge in Natural language processing, Deep learning, Spark, PyTorch, Transformers, RoBERTa, Designing Machine learning experiments.
Co-founder/Tech lead
inVOID Technologies
June 2018 – May 2019
• Cofounded tech startup and lead product development while acquiring investors and customers.
• Lead the entire technical efforts of the startup and developed POCs and Pilots for our ID Verification prototype.
• Researched and Developed facial recognition models achieving state of the art accuracy of >99%.
• Created novel liveliness detection software system for ID verification and fraud prevention.
• Leveraged knowledge in Python, Deep learning, Machine learning, Computer Vision, TensorFlow, PyTorch, RESTful APIs, ReactJS, JavaScript, Flask, MongoDB, NoSQL, Redis, Docker.
Research Intern
Complex systems lab IIIT Delhi
June 2017 – May 2018
• Created FlavorDB, the most extensive bioinformatics database for flavor molecules, and their properties, published in 2018 issue of Nucleic Acids Research journal a very high impact journal, gaining more than 26 citations under a year.
• It is now being integrated into amazon Alexa to power its food pairing recommendations.
• Developed full stack web applications with data visualization using D3.js on the front end and performant backend REST APIs in flask.
-
Projects
• Recommender System, INF553: Created Item based, model based, and hybrid recommendation system for recommending restaurants using yelp data. Tech stack: Python, Spark, XGBoost
• Comic Strip Generation, CSCI566 Deep learning: Created novel ways of generating Garfield comics using deep learning. Which included GANs, Seq2Seq (LSTMs), Multimodal approaches. Tech stack: Tensorflow, Pytorch, Python
• Halma game-playing agent, CSCI561 Artificial Intelligence: Created an AI agent using Minimax and alpha-beta pruning to play a game of Halma(Chinese-checkers), beating 80% of other submission. Tech stack: Python