At Deloitte USI, I architected a supervisor-agent system for a leading US bank, with privilege-scoped sub-agents and dedicated tool calls. The system serves 4,000+ users and is in full-scale production rollout across the bank’s wealth management business.
I engineered grounded prompts and RAG over CRM data, and built feedback and observability for model generations. I also re-architected a critical service, reducing processing time from 14,000ms to about 300ms.
Outside work, I independently built multi-agent products including an autonomous trading platform with a deterministic risk engine and an AI SDLC pipeline grounded in real codebases. I work daily with remote, cross-functional stakeholders across time zones to turn requirements into shipped software.
