Muzammil Mohammed

I'm building and deploying advanced voicebots at EXL using Playbooks and GECX Agent Studio, automating complex customer-support workflows while reducing AHT and operational overhead. I also create agent parameters and logging tools that make complex conversation flows easier to debug.

I developed a Condition + Action prompting technique that improved model reliability by 70% through reduced hallucinations and overfitting. I redesigned prompt architectures for customizable barge-in and multilingual support, and load-tested voicebots at 1,600+ concurrent calls.

At Accenture, I developed T-Mobile chatbots and voicebots with Dialogflow CX, Google Speech, Generative Playbooks, Cloud Functions, external APIs, webhooks, databases, and messaging queues. I also built Agentic AI workflows with LangGraph and Gemini 2.0 Flash that increased playbook creation efficiency by 40% and cut development time from two weeks to one.

Previously at Foundation AI, I built queue-based Python pipelines processing millions of PostgreSQL records, reducing latency by 70%, and improved prompt quality while lowering token costs by 12%. I bring a Responsible AI mindset to conversational design, guardrails, RAG, NLU optimization, UAT, and production hypercare.