At Tata Consultancy Services, I engineered an autonomous Agent and MCP Server for Lloyds Banking Group to retrieve, consolidate, and summarise complaint data across 12 enterprise systems. It reduced complaint effort by 97% and average handling time from 60 minutes to 2 minutes.
I also contributed to HIVE, an enterprise multi-agent platform, building four reusable AI agents and tools for colleague workflows. I developed production-grade agent workflows with groundedness evaluation, tool sanitisation, guardrails, and end-to-end observability.
As an AI Intern at Accenture, I built an LLM-powered application for Honda to query project databases and access production information and insights. It integrated data summarisation, anomaly detection, and multilingual support.
At RenderDash Studio LLP, I co-founded the company and built an AI analytics platform that generates dashboards, visualisations, and forecast tables from sales data, with demand forecasting at 92% accuracy. My projects include a CrewAI-based landing page generation system and multi-agent orchestration pipelines with parallel execution and adversarial verification.
