Saipavan Singala

Ai Machine Learning Engineer @PNC

US

Developed and fine-tuned transformer models such as FinBERT, GPT-3, and LLaMA for financial sentiment analysis, fraud detection, and risk modeling, achieving an F1-score of 0.92 and enabling real-time analysis of unstructured financial data. Built anomaly detection systems using Autoencoders and Graph Neural Networks, improving fraud detection accuracy to 93% and reducing false positives in compliance monitoring, deployed on AWS SageMaker for real-time inference. Designed retrieval-based AI systems leveraging Retrieval-Augmented Generation (RAG) to automate regulatory reporting, extract insights from financial filings, and enhance investment research workflows. Implemented end-to-end ML pipelines with MLOps best practices, including CI/CD, model monitoring, and scalable deployments on AWS and Google Cloud AI, ensuring high system reliability for trading and compliance use cases. Established AI-powered chatbots and voice assistants for wealth management and client services, improving query resolution speed and enhancing personalized investment recommendations. Applied reinforcement learning for trading strategy optimization and portfolio allocation, and leveraged time-series forecasting models to predict market trends for investment strategies. Optimized ML workflows using Apache Beam, Spark, and Dataflow to handle large-scale financial datasets, while deploying containerized models via Docker and Bitbucket CI/CD pipelines.

Lovely Professional University (LPU)

Bachelor's degree, Computer Science and Engineering

University of Bridgeport

Master's degree, Computer Science