Worked as a member of the ML Growth Team in different capacities and projectsContributed to enhance relationship detection system using heuristic and machine learning from company homepages, by making it robust by improving its RESTAPI, system monitoring, and enhancing the analytical tracking of the overall system.Designed and contributed to developing event-driven system that reacts to changes in risk scores or user activities. This system is designed to automatically identify and extract relevant business relationships from corporate websites.Contributed to research and developing a proof of concept to detecting products and services from raw unstructured text and linking them to different classification ontologies through deep learning method as part of ML discovery team.Contributed to designing and developing a ML based entity resolution service using transformer, vector databases to resolve entity from a knowledge base of hundreds of millions of entities.

Developed Human in the loop feedback system to measure the performance of the NLP models, detect concept drift, model drift and build new training datasets.Iteratively improve model performance by re-training event-detection and name entity recognition models for low-resource business eventsImproved REST API design and implementation from NLP models inferencesDesigned and developed python packages for data preprocessing and data schema to store and read datasets.Researched and reimplemented academic research paper as part of the solution for event detection models using generative adversarial networks in NLP

Member of Machine Learning (ML) Information Extraction team building statistical and deep learning models to extract information from unstructured data sources.My work focuses specifically on building scalable models to solve problems like entity linking, entity extraction, and event detection. Building REST APIs for trained ML models, designing feedback pipelines for continuous improvement of these ML models in production.I have also worked on a computer vision problem for building a scalable search engine with an active learning component.

I worked as a research assistant in the Ebiquity Lab under Dr. Francis Ferraro. My research involved work on knowledge representation and its implications in improving downstream natural language processing applications like question answering system, language modeling.

Developed information extraction tools for mapping open web data to the knowledge base to enhance the data ingestion pipeline.Build an intelligent product crawler that uses both unsupervised and supervised machine learning to learn the HTML structure of the webpage and identify the product names accurately on those webpages Developed a pipeline for intelligent product crawler tool to validate and map the information to the organization entities in the knowledge graphCollaborated with Senior Data Scientist in building statistical machine learning to measure the relationships and operations risk health factor in a supply chain network

• Designed and developed a deep learning algorithm that predicts the patient’s need for massive transfusion within 1st hour and first 24 hours post-trauma by studying the first 15 mins of PPG signal • Developed a multi-task machine learning algorithms by studying vital signs of trauma patients to predict 7 different outcomes and discovered relationships between the outcomes through different statistical regularization techniques. • Collected real-time patient data from a database of size 150TB and prepared datasets for machine learning algorithms

Worked and collaborated with the cross-business unit teams in developing python libraries for robotic process automation applications. Programmed shell scripts for automating daily activities related to operating system and Database. Conducted research with an interdisciplinary team of the database, operating systems and networking experts to optimize the performance of SAP data migration across data centers. Worked with solution architects in designing end-to-end implementation and configuration of SAP applications in complex environments. Enhanced SAP application capabilities with version upgrades, system refreshes and add-on installations. Provided operational support to client SAP environment, devised strategies for integration of SAP Infrastructure with add-on non-SAP tools, developed performance tuning analysis to prioritize the changes in the client’s SAP infrastructure to avoid application failure.