Pugazhendhi Rathinasabapathy

Ai Ml Engineer @Northern Trust

Boston, MA, US

Spearheaded the development of a machine learning platform using Python, Scikit-learn, XGBoost, Regex, and DistilBERT to model portfolio risk categories, improving risk detection accuracy by 30%.• Engineered advanced financial features such as asset correlation, sector exposure, and volatility from market data using Pandas, NumPy, SQL (Snowflake), and Word2Vec, boosting predictive power by 25%.• Built and optimized deep learning models in TensorFlow integrated with Large Language Models (LLMs) using LangChain, LlamaIndex, Groq API, and RAG frameworks to capture complex asset behaviors, enabling proactive risk scenario analysis and reducing portfolio volatility.• Developed an automated Apache Airflow pipeline for daily data refresh, model retraining, validation with Pydantic, and AWS deployment on EC2 and S3, delivering real-time risk outputs with 40% faster latency.• Implemented SHAP explainability to enhance model transparency, increasing advisor trust and meeting audit standards.• Delivered interactive Power BI dashboards visualizing portfolio risk trends and AI-driven reallocation strategies using vector search technologies FAISS and Pinecone.• Collaborated cross-functionally with data engineering, investment, and compliance teams to align ML solutions with regulatory requirements, ensuring scalable and audit-ready AI systems.