Data science master's student at Harvard University and Oxford mathematics graduate with First Class Honours. Experience in machine learning, quantitative modeling, SQL database design and mathematical evaluation of large language models.
Overview
2
2
Languages
1
1
year of post-secondary education
2
2
years of professional experience
Work History
Math Expert
Outlier, Scale AI
09.2024 - Current
Authored challenging mathematics problems with structured LaTeX solutions and evaluated model responses, enhancing training for large language models.
Reviewed step-by-step reasoning for mathematical accuracy and clarity.
Reviewed and assessed math solutions for Outlier and Scale AI model training.
Annotated problem sets using project guidelines and quality standards.
Mathematical Engineer
Quant
02.2026 - 03.2026
Developed functional mean reversion concept using symbolic regression to identify multivariate asset-price functions maintaining stability during specific intraday windows.
Developed Quant engineering solutions for trading and data platforms.
Refined Quant system requirements with product and operations teams.
Implemented automation scripts for Quant deployment and monitoring tasks.
Student One Causal Networks Pvt Ltd
Education
Master of Science - Data Science
Harvard University
Boston, US
MIT (Cross-registration) -
MIT
Cambridge, US
Certificate - Quantitative Finance
CQF Institute
London , UK
01.2025 - 08.2025
Bachelor of Arts - Mathematics
University of Oxford
UK
Skills
Machine learning
Statistical modeling
Quantitative analysis
ML and AI: MLOps, LLMOps, RAG techniques, LLM expertise
Tools and platforms: Git, GitHub, Jupyter, Docker, Kubernetes, GCP
Personal Information
Title: DATA SCIENCE | MACHINE LEARNING | QUANTITATIVE MODELLING
EXTRACURRICULAR ACTIVITIES AND INTERESTS
Chess | Music: Guitar playing | Fitness: Strength and weight training
SELECTED PROJECTS
Etsy Stock Return Prediction | Machine Learning Ensemble View GitHub repository, Built a binary classifier to predict whether Etsy's 15-day forward return would exceed 5%, using 2015-2025 daily OHLCV data and S&P 500 features., Removed highly collinear features, applied Standard, Robust and MinMax scaling, and used KMeans to select the feature nearest each cluster centroid., Tuned nine base classifiers with TimeSeriesSplit and ROC-AUC GridSearchCV; combined them in a two-layer ensemble achieving 76% test accuracy and 0.70 AUC.
SQL Meal Tracker Database | SQLite View GitHub repository, Designed a five-table relational database to track ingredients, meals, consumption logs and daily macronutrients, using composite keys, foreign keys and constraints for reliable CRUD operations., Implemented joins and subqueries, automated daily nutrition aggregation with SQL triggers, and indexed frequently queried fields to improve lookup efficiency.
JPMorgan Chase & Co. Quantitative Research | Forage Virtual Simulation, Modelled natural gas prices with month-wise linear regression to price commodity storage contracts; used logistic regression to estimate loan default probability and expected loss., Applied dynamic programming to bucket FICO scores into categories that maximise log-likelihood.