Experience

Artificial Intelligence Intern | New York, NY | May 2025 - September 2025

Built a Python ML framework for preprocessing, training, evaluation, and artifact management, helping 22 data scientists across 4 teams launch experiments in 10 minutes instead of 2 hours.
Integrated MLflow for automated tracking across 100+ training runs with AWS S3 artifact storage and role-based access control.
Reduced model handoff time from 1 day to 5 minutes while managing 50GB+ of model artifacts.

Software Developer | Los Angeles, CA | December 2024 - May 2025

Built a Selenium scraping pipeline with checkpoint recovery and rate limiting to process 180,000+ filings.
Reduced end-to-end scraping time from 5 days to 2 days through full automation.
Deployed a 4-model LLM ensemble (Claude, GPT, Llama, Gemini) with cross-model validation to classify 72M+ words into 26 categories.
Engineered scalable pipelines for 100,000+ abstracts and journal metadata to support retrieval-augmented generation workflows.
Augmented data 10x by generating 500+ abstracts with OpenAI APIs for RAG model fine-tuning in scientific literature retrieval.

Data Science Intern | Mumbai, India | June 2023 - October 2023

Programmed a data augmentation workflow with SDV for low-sample scenarios, improving model accuracy by 13%.
Built customer churn and engagement models with Scikit-Learn and XGBoost, improving recommendation and decision workflows by 9%.
Developed a web-based sensitivity analysis tool using React, FastAPI, and PostgreSQL used by 15+ teams.

Software Engineering Intern | Mumbai, India | December 2021 - July 2022

Developed cross-sales filtering to forecast customer trends, increasing targeted product revenue by 15%.
Integrated real-time APIs with credit scoring services to build personalized financial profiles and improve recommendation quality.
Built a React + Laravel + MongoDB proof-of-concept for migration from a PHP stack, improving load time and user feedback by 20%.

Software Engineering Intern | Mumbai, India | May 2021 - August 2021

Built a React + Flask application to detect underperforming APIs and improve backend performance by 39%.
Deployed Flask services on AWS EC2 with Auto Scaling Groups and an Application Load Balancer, scaling from 2 to 10 instances for 7,000+ concurrent requests under 200ms latency.
Used OpenCV quality checks to detect broken tablets and color mismatches using strict thresholds.