My Monologue

Hi, thanks for checking in! I like going on random tangents when I write, so I’m not sure if this is the best way to write about myself. But hey, I’m glad you’re here.

I have enough places to tell you about what I have been up to, so let me take some random tangents about me here (do check out my blog for worklogs if you’re interested though!).

I studied Computer Science Engineering at Shri Mata Vaishno Devi University in Katra, India, graduating in 2023. Somewhere in the middle of that, I did a stint with Google Summer of Code under the Python Software Foundation’s SUSE EOS organization, building product-tracking and community-management features in JavaScript. That was my first real taste of shipping code that other people actually used, and it stuck.

After graduating, I joined Tata Elxsi as a Software Developer, where I ended up doing a little bit of everything: an AI-driven unit test case generator using CrewAI and Gemini, UML diagram generation from automotive feature specs using Claude on AWS Bedrock, a RAG-based PDF chatbot with Mistral 7B and Streamlit, and a LangGraph-based agentic system for multi-cloud cost governance with anomaly detection and Slack-based human-in-the-loop approvals. Somewhere in there, I picked up an Extra Mile Award for the test automation work and a customer appreciation award for a secure RAG pipeline I set up in the cloud.

That stretch of building agents and RAG systems for real production problems is what pulled me deeper into GenAI specifically, rather than software engineering more broadly. So I moved to Sasken Technologies as a Senior Software Developer, where I now build agentic systems in earnest — a predictive maintenance system with three specialized LangGraph agents (root cause analysis, planning, and scheduling) that diagnose and resolve faults across motors, batteries, and gearboxes, along with a conversational AI shopping assistant for natural product discovery.

I like working at the intersection of “this needs to actually work in production” and “this is genuinely a hard agentic-systems problem.” Multi-agent coordination, retrieval, and getting LLMs to behave reliably under real constraints are the things I find myself gravitating toward, and I expect that’s mostly what you’ll find me writing about here.