-Worked on Cohere’s core platform on the intake team, supporting the processing of 275k cases annually, driving efficiency and reducing manual interventions-Researching and implementing machine learning models for prediction of medical procedure codes, optimizing decision-making processes - Implemented and released over 20 front-end and back-end features, leading to a 40% reduction in production bugs through rigorous end-to-end Cypress testing- Developed an ETL pipeline using Kafka for real-time data ingestion and Apache Hudi for incremental storage, enabling fast queries and real-time analytics with seamless historical data backfill to an S3 Bronze Table- Worked with Python, PyTorch, React.js, Java, MongoDB, AWS Tools, Kafka and Cypress

- CS381: Analysis of Algorithms- Responsible for organizing the course, managing logistics, and developing content for a class of approximately 430 students- Held office hours to assist students in understanding algorithmic design principles

- UTA for CS251: Data Structures and Algorithms- UTA for CS381: Analysis of Algorithms - Conducted weekly office hours, providing support to students, addressing their questions, and clarifying course concepts.- Graded assignments and provided constructive feedback to students, aiding their understanding of course materials.

- Developed a robust Flask REST API to integrate a pre-trained Multi-Label Classifier in a system withmicroservice architecture enhancing overall functionality- Integration of API optimized data flow and improved processing with a 10% increase in efficiency- Deployed API on AWS EC2 instances using Docker Container to ensure scalability and high performance- Developed in Python using scikit-learn, Keras, pandas, nltk, Flask and Docker

Project: Control My News Feed: Exploring Users Manipulate Social Media Recommendation Systems.Researching social media addiction to build better recommender systems.

Project: Investigating Explainability of Machine Learning Model Decisions for Injury Surveillance.Built ML Models with 92% accuracy for predicting injury narratives in the workplace.Created an explainable interface to explain the model decisions and improve upon its accuracy throughhuman coding by 5%.Developed in Python using tensorflow, keras, scikit-learn, LIME, sHAP and InterpretML

- Created Data Pipelines using web scraping frameworks to consolidate datasets with over 50,000 data points from websites like Amazon, Reddit and Twitter related to product usage and reviews- Worked in a team of 4 to develop Natural Language Processing (NLP) Models with 85% accuracy identifying injury cases, categorizing body parts and classifying injury cases- Deployed the NLP Models as a REST API within a Docker container to enable real-time predictions- Developed in Python using tensorflow, scikit-learn, Keras, Flask, Selenium and Docker

- Created a Java Application with a team of 6 to process and restructure 1GB of raw data for analysis usedfor preparing client reports- Developed a REST API using Spring Boot that implements an efficient API endpoint that responds torequests with serialized data in JSON format- Tracked and documented bug fixes using JUnit test cases increasing software stability by 15%- Developed in Java using Spring Boot, SQL, JDBC, JSON and MS Access