At Pegasystems, I’m a core backend engineer on Pega Autopilot, an agentic AI framework now generally available across the Pega Infinity platform. I implemented multi-provider model orchestration across OpenAI, Amazon Bedrock, and Google Vertex AI, enabling new models to go live at runtime without a new Infinity release.
I designed and built agent architecture, templates, and execution loops with tool calling, conversation management, context compaction, and prompt caching, cutting per-conversation token spend by 30%. I also created evaluation infrastructure with DeepEval, onboarding 50+ agents from 10+ internal teams to automated testing and monitoring.
For Pega Autopilot, I engineered a RAG pipeline using Document AI and Amazon OCR, and fine-tuned models with Pega-specific datasets. I also authored a Gradle plugin that made agent regression testing a per-release gate.
Before Pegasystems, I worked on production NLP, IoT device management, automotive testing, and e-Government platforms at Ericsson Global Services Pvt. Ltd., Mercedes-Benz Research & Development India, and CrimsonLogic Pte. Ltd. I’m currently pursuing an MTech in Artificial Intelligence and Machine Learning at BITS Pilani.
