Master of Professional Studies in Data Science

GPA: 4.0

Bachelor of Technology in Information Technology

Senior Research Assistant

Automated Patient Interview Analysis Tool (Mellowing Minds Study)
● Designed and implemented software tool to infer stress reduction ways for corresponding medical symptoms by applying
k-means clustering with BERT model on interview data, involving 200 patients study data
● Developed a module for treatment assessment using feature extraction and topic modeling for 500+ patients’ illness categories
● Monitored performance by conducting hypothesis testing and correlation analysis improving patient recovery time by 55%

Associate Data Scientist

B2B Content Intelligence Advanced Recommender Systems
● Leveraged NLP techniques and Elasticsearch to engineer hybrid recommender systems, leading to 200% increase in site traffic
● Created and Integrated data processing modules using techniques like word2vec, BERT, LDA and TF-IDF, and built algorithm
selection tool to optimize performance, achieving 70% improvement in content relevance
● Implemented scalable recommender API as a microservice enabling multi-client compatibility through containerization with
Docker and orchestrated deployment using K8, revenue exceeding $20 million for each client

Systems Engineer

BaNCS Swift Messages Interface
● Built sustainable Interfaces (Position & Trade) for CitiBank & Société Générale on Java, with capacity load of 100k trades per day
● Played a key role by collaborating to develop Unix scripts with DevOps team, executing Oracle DB migration for 300+ tables
● Executed back-end rules with Java and RabbitMQ for secure tunnel processing of SWIFT messages and RIC & REIT securities

Software & Data Science Intern

Punjabi Information Report Insight Engine
● Collaborated with Punjab police to produce software that analyzes Punjabi language FIRs using OCR Tesseract, identifying
high-risk zones and aiding in a 50% accident reduction
● Led a team to automate the process of determining the cause of FIRs using NLP and Punjabi embeddings, achieving an analysis
rate of 100 FIRs per hour