Ishwar G.

About Ishwar G.

I am ML Engineer with experience in machine learning, deep learning, and natural language…

Experience

1. Ai Ml Engineer

   Unum

   Feb 2024 — Present · US

   Designed RAG-based LLM workflows with LangChain and vector embeddings, enabling claims teams to query policy documents in natural language, cutting manual search time by 60% and driving $3.2M in annual savings.• Engineered Fast-API-based services for insurance claims decisioning, integrating ML models with distributed microservices on Azure ACI to serve 50K+ daily requests with 95% of request less than 180ms latency.• Built a GenAI-powered summarization service with Hugging Face and PyTorch to auto-generate claim dispute summaries, reducing the case resolution cycle by 3.5 days and directly improving customer satisfaction.• Designed and implemented a PyTorch QLoRA fine-tuning pipeline for domain LLMs, boosting vector-database retrieval-filter accuracy by 35% and cutting compute costs by 60% for insurance-claims workflows• Developed vLLM based inference and deployed ML services on Azure AI (ACR, ACI), enabled continuous batching/PagedAttention to lift throughput 45%, with GitHub Actions CI/CD for reliable releases.• Implemented ML monitoring on Azure with MLflow, Prometheus and Grafana, adding drift alerts and auto rollback and reducing incidents by 30%.• Partnered with data engineers to create a Snowflake and dbt feature store refreshed daily, ensuring underwriters accessed the latest claim and policy signals for pricing, reducing error rates by 22%.

Education

• Dr. Babasaheb Ambedkar Technological University

  Bachelor of Technology - BTech, Computer Engineering

• The University of Texas at Dallas

  Master's degree, Business Analytics

• The University of Texas at Dallas

  Graduate Certificate, Applied Machine Learning