Will Po

Pioneered and built agentic and generative AI at Woolworths, including the architecture and delivery of its conversational platform. Delivered production agents such as a Loyalty Rewards Agent on Google CX Agent Studio (CXAS) with Python tool/functions, end-to-end evaluation, operational monitoring, and an MCP Proxy approach to add/remove MCP servers into agent context without redeployments. Shipped a customer-facing Shopping Agent using a custom TypeScript harness (Gemini) with context management, deterministic fallbacks, real-time guardrails, and evaluation; implemented RAG with Qdrant and model routing/fallback strategies, optimising for low latency and reliability. Took early GenAI contact-centre initiatives to production with evaluation frameworks and guardrails, reducing post-interaction admin time from ~5 minutes to ~30 seconds and saving $1M+ annually. Led an Olive Customer Chatbot conversion to an agentic assistant with CXAS, achieving close to 90% contact-centre deflection a

Spent a decade as a senior engineer embedding with engineering and product teams across retail, media, financial services, airlines, insurance, and technology startups. Delivered end-to-end React and Node.js applications with unified API/GraphQL layers and enterprise backend integrations. Built and maintained scalable serverless systems on AWS and GCP with automated deployments, often joining mid-flight projects to recover stalled delivery and ship stable production systems.

Worked across retail, media, financial services, airlines, insurance, and technology startups, joining projects mid-flight and delivering quickly. Delivered end-to-end React and Node.js applications with unified API and GraphQL layers, plus backend integrations across enterprise systems. Built and maintained scalable serverless systems on AWS and GCP with automated deployments, recovering stalled delivery timelines and shipping stable production software.