Anjali Pugalia

I’m a Data Scientist with 3 years of production experience designing quantitative models and AI/ML/NLP/LLM/RAG-based analytical solutions in financial services. I focus on taking ideas from proof-of-concept to production with measurable business outcomes.

At Wells Fargo, I designed and deployed a real-time text classification model for operational risk tagging, integrating it with Kafka and deploying on Pivotal Cloud Foundry. The work automated manual review and delivered ~$100K in annual savings.

I also built a scalable Nearest Neighbors semantic search engine using Word2Vec embeddings and SQL-based preprocessing, deployed on Teradata for high-volume similarity-based text retrieval. That effort drove ~$200K in annual savings through improved data quality and decision support.

I’ve built and iterated on GenAI solutions as well—developing a RAG-based chatbot proof-of-concept with fine-tuned LLMs on domain-specific financial data, and designing an LLM-powered virtual assistant for real-time complaint capture and classification via REST API integrations. Across these projects, I collaborate closely with business, product, and engineering stakeholders to deliver scalable, production-ready systems.