Jan 2024 – Present · 2 yrs 2 mos · Bengaluru, Karnataka, India · On-site

• Enginnered the modernization of legacy systems and contributed to the development of multi-bot LLM-based ticketing systems, resulting in over 95% task automation accuracy.
• Engineered Agentic AI workflows leveraging LangGraph for SQL error detection, integrating with MS Teams and ServiceNow to facilitate automated ticketing and human-in-the-loop review processes.
• Implemented Retrieval-Augmented Generation (RAG) pipelines utilizing Pinecone, Azure Cognitive Search, and LLMs to automate document understanding and insights generation across various client projects.
• Developed a layout detection module using Python, OpenCV, and Tesseract OCR to convert scanned documents into structured JSON (pages, headers, tables, figures), leading to a 13% improvement in extraction accuracy.
• Constructed NER-based and decision-tree models using TensorFlow for automated field extraction and validation, resulting in increased data accuracy and reduced reliance on rule-based dependencies.