1. Lead a group of 10 students with varying levels of experience to create a solution to predict the experimental conditions for HPLC-MS/MS Assay for Pharmacokinetics using NLP. This setup will reduce the trial and error needed to perform tests on new drugs and will speed up the drug delivery process.2. Worked with Bio-analysts from Merck, the key stakeholders, to identify pain points in their process as well as gain subject knowledge to design the right solution for them.3. The project was organized into four components: a. Building an optimized data pipeline to extract high-quality HPLC-MS/MS abstracts from PubMed b. Manually annotating abstracts to create a data corpus to train an NLP Model c. Training, evaluating and comparing multiple NLP models (pre-trained and un-trained) to extract assays from abstracts with high accuracy. We focused on vanilla BERT, BioBERT, spaCy, and RoBERTa. d. Easy-to-use website built for Bio-Analysts to query the curated assays to find their information quickly and efficiently

- Designed and built a Fan-out Fan-in architecture using AWS services like Lambda, S3, and DynamoDB to automate the ingestion process in the ETL Pipeline to load a large dataset from a PostgreSQL data source into a Data Lake on Databricks- Reduced the time consumed for ingestion from 4hrs to 15min and allowed data scientists to perform analysis and provide key insights to help customers improve efficiency and reduce costs- Followed CI/CD procedures with Jenkins, wrote Infrastructure as Code with Terraform scripts and, packaged and versioned code in containers using Docker and AWS ECR

Guided students in the Introduction to Python Programming and Machine Learning course at Purdue by conducting office hours and grading assignments

Evaluated and guided students in the Introduction to Digital Systems Design (ECE 270) course at Purdue

Biometrics Project with Merck:- Developed a full-stack web application on Flask to communicate biometric data collected from wearable fitness technology for determining the effect on mental and physical performance during the clinical trial phase of a new drug lifecycle.- Designed and lead the development of a mobile application on React Native with a Node REST API server to collect qualitative information from patients and a dashboard on RShiny- Served as a development lead and coordinated tasks among 15 students with varied experience on the team

- Responsible for building optimal scheduling systems, solving constraint satisfaction problems, building scalable architectures and creating a complete automated test suite.- Designed and developed a new interface for businesses on a full-stack web-based platform with core components consisting of a React with Redux frontend, Node JS REST API server, Google OAuth, MySQL database, and Google Cloud- Created a complete suite of automated unit and integration tests for the API endpoints using the Mocha testing framework with the Chai assertion library- Built End-to-End tests for all user stories on the frontend using Selenium Web-driver and CyPress

Supervised and guided the practical work of students by providing them timely and effective feedback to support their learning in the course PHYS 172.