Vasanthi N.

About

ML/NLP practitioner focused on financial-services document intelligence and compliance workflows—built an end-to-end pipeline to classify documents and extract financial entities from loan applications, emails, and statements stored in S3/internal databases. Strong in entity resolution/record linkage and in productionizing pipelines with GitHub Actions CI/CD, testing, data validation, and Docker, plus semantic search using OpenAI embeddings and a vector database.

Experience

Education

Key Strengths

• Built end-to-end NLP pipeline for unstructured financial documents (loan applications, emails, financial statements)
• Entity resolution across emails, loan documents, and CRM data using blocking + similarity + classification
• Model selection balancing interpretability (logistic regression) vs performance (XGBoost)
• Validation approach combining manual audits with business analysts and post-deployment monitoring
• Productionizing ML/NLP workflows with CI/CD (GitHub Actions), automated tests, data validation, and Dockerized deployments
• Prototype-to-production decisioning using precision/recall and runtime on labeled real-world datasets