Associate Product Developer with expertise in BMC Adapt-Angular, FastAPI, and RAG-based systems. Proven track record in delivering enhanced UI components and AI-driven applications
Overview
3
3
years of professional experience
Work History
Associate Product Developer
BMC Software
Bangalore
07.2022 - Current
Built and enhanced UI panels using the BMC Adapt Angular framework, including implementing Adapt tables (covering approximately 80% of the product structure) and enabling lazy loading across the IMS Command Center, which significantly improved performance. Recognized and appreciated by product managers and senior leadership for these contributions.
Created an intuitive, modern JCL editor web component with real-time error detection and advanced features, soon to be released as an npm package under bmc-des for wider use. Appreciated by senior developers and the manager.
Developed a RAG-based application using FastAPI, integrating vector databases like FAISS, with LangChain and LangGraph, to deliver accurate, domain-specific responses.
Added a GPT-style chat interface within the IMS Command Center using Adapt Angular, integrated with RAG pipelines, and tool-calling for Agentic Flow, enabling the system to retrieve live IMS data. Presented to product line operations, and the prototype is moving toward full product integration.
Created an AI-driven analysis feature for IMS Message Advisor queues by retrieving real-time data from the mainframe via PCH/UIM APIs, combining multiple system prompts with RAG strategies. It is also moving forward with product line operations.
Explored and implemented multiple chunking strategies for RAG during a Root Cause Analysis hackathon, building an interactive dashboard to demonstrate the workflow. Conducted demos and knowledge-sharing sessions with product teams, and it is moving toward full product integration.
Designed and built RAGMaster, a reusable project that allows users to query their own document collections with full RAG and Self-RAG support, backed by detailed documentation and algorithms.
Authored a research paper on agentic AI and RAG implementations in enterprise systems, outlining technical details, architectural approaches, and a complete implementation guide.
Pre-trained custom small language models (15M and 20M parameters) with a 256-token context window on an NVIDIA RTX 3070 Ti (8GB), inspired by OpenAI’s TinyStories research. Completed 3-epoch training in approximately 13 hours, capable of generating coherent short stories.
I developed an NLP-powered chatbot during my final-year internship at Celebal Technologies, using TensorFlow for a client project.
Awarded a Gold Medal in IBM ICE for a project on creating VSAM and GDG datasets through Jenkins pipelines.