Roland Taylor

Technology Engineering Lead, AI and Automation, at Workwear Group (Wesfarmers) in Melbourne.

Agentic engineer who builds AI agents and tools with Claude and AI coding tools, then sells the idea to the people who fund it. Replaced an outsourced IT service desk with an AI agent at about 2% of the cost, runs discovery workshops and AI training for leadership, and builds working demos fast enough to change a decision in the same meeting. Background across telco, aviation, retail, energy and apparel, from the sales floor to enterprise architecture.

Quick answers

Agentic engineering
Builds and runs AI agents on Azure (Nova, Sage, Safety Sam, Pierce, PAUL), a CRM and this CV’s AI avatar with Claude Code and other AI coding agents.
Pre-sales
Sold on a retail floor, pitches his own builds to the CIO and executives to win their funding, and has sat on the buyer’s side of many vendor pitches.
Discovery and demos
Ran discovery from account managers to executives before building the CRM, built an agent live with a leadership room as the client, and built an agentic shopping demo on three brands’ storefronts.
Business cases
Nova runs for about $600 a month against the $30k a month of the desk it replaced, with CSAT at 96%.
Salesforce
Hands-on with B2C Commerce: rescued Apple Pay across 10 merchant accounts by migrating it to CyberSource’s REST cartridge before the certificates expired.

Experience

2022 to now

Workwear Group (Wesfarmers), Melbourne

Technology Engineering Lead, AI and Automation, from October 2025Technology Lead, September 2024 to October 2025Senior IT Engineer, Identity and End User Computing, June 2022 to September 2024

  • Replaced a $30k-a-month outsourced service desk with Nova, an AI agent that is the IT service desk for about 500 staff. Nova runs for about $600 a month, dev and prod together. CSAT has held at 96% while the number of ratings rose about 400%.
  • Designed, built and runs the company’s AI agents on the group’s Enterprise Landing Zone on Azure, each with its own landing zone, CI/CD, sign-in and monitoring: Nova (IT service desk), Sage (HR policy), Safety Sam (safety reporting, live to all employees since October 2026), Pierce (application support) and PAUL (Jira and Confluence).
  • Built ACSEA, AI customer service email automation that handles more than 700 emails a day, on Logic Apps and Azure Service Bus.
  • Ran discovery with account managers, their managers and executives, then built a bespoke wholesale CRM with AI coding tools: Next.js, PostgreSQL with pgvector, Azure OpenAI and Microsoft Graph webhooks for email. About $140 a month, with no per-seat licences.
  • Rescued Apple Pay across the company’s Salesforce B2C Commerce storefronts (Hard Yakka, KingGee and NNT, AU and NZ, 10 merchant accounts). A routine certificate renewal exposed that the legacy CyberSource SOAP cartridge no longer supported Apple Pay with new-format keys. Diagnosed it from the SFCC pipeline logs, got CyberSource to confirm the root cause, and migrated Apple Pay to their REST cartridge before the certificates expired.
  • Built an agentic shopping demo across Hard Yakka, KingGee and NNT on Azure AI Foundry, indexing their Salesforce B2C Commerce catalogues, with a visible rail of the agent’s reasoning, photo-to-product matching and a kit builder. Built for the AI Community of Practice and leadership.
  • Wrote “Customer Service, Our Way”, the proposal for one AI agent across every brand and channel, grounded in analysis of Freshdesk tickets, with an MCP tool layer into the IBM i ERP.
  • Built a contract intelligence pipeline that pulls contract documents into Snowflake and reads them with AI_EXTRACT.
  • Argued against about $250k of vendor pitches for front-line AI sales agents: the money was better spent on being recommended by, and buyable through, AI assistants. Pitched an AI crawler accessibility strategy to Digital and Marketing leadership.
  • Technical lead and decision-maker for the company on an engagement co-funded by the group’s AI Accelerator and Microsoft. It delivered the Enterprise and AI Landing Zones, and IBIS, an excess-inventory agent, is in flight.
  • Packaged Nova for the group’s AI Accelerator marketplace so other divisions can reuse it. It was the first submission that was a working agent.
  • Runs the AI Centre of Excellence and Community of Practice. Designed and taught the RISE Leadership Academy’s Digital and Data workshop solo in October 2026, where the room briefed an agent and watched it built live. Built Prompt Lab, a prompt-engineering academy where staff pass by prompting a real AI that grades them, and a SCORM e-learning module on responsible AI.
  • Final escalation point for the engineers in Cyber Security, End User Computing, Cloud and Infrastructure, and AI and Automation.

2018 to 2022

Origin Energy, Melbourne

Senior IT Engineer, October 2019 to June 2022Project Coordinator (contract), from June 2018

  • Brought in on a contract day rate to merge three Active Directory domains into one, for about 8,000 staff across continents.
  • Moved the Citrix estate from on-premises to AWS, then on to Azure Virtual Desktop to support 2,000 staff in Manila. That second move was about 80% done when the pandemic hit.
  • When the pandemic hit, scaled the remote-access platform to three times its size while keeping the offshore call centres running. Within a week all 8,000 staff in five countries could work securely from home, including at drilling sites and power plants on operational technology networks.

2016 to 2018

Kinetic IT

Senior Service Desk Technician

  • Started on the phones on the Qantas service desk and made senior within a year. Moved to Melbourne for the Origin Energy account, and was hired away by Origin within a year.

2015 to 2016

Harvey Norman

AV and IT Senior Sales Consultant

  • Regularly hit stretch targets and maxed out the monthly incentives selling AV and IT, with repeat customers and referrals by name.

2014

Airbnb

Host Activations

2013 to 2014

iiNet

Service Desk Technician

  • Moved from the service desk to the team that handled cases escalated to the Telecommunications Industry Ombudsman.

Skills

Claude Code, agent loops and tool use, MCP servers, context engineering and evals · Azure (Landing Zones, AI Foundry, Logic Apps, Service Bus) · Microsoft 365 and Copilot Studio · Snowflake · Power BI · Next.js and PostgreSQL · identity and OAuth · Salesforce B2C Commerce (payments, cartridges, Apple Pay, storefront catalogues)

Education

2009

The New Zealand Film and Television School

2020

AWS Certified Solutions Architect, Associate

Outside work

Runs SOL, a home server of about 65 containers kept as code and deployed through CI. My AI avatar runs on it, alongside an MCP server that gives Claude operational access and a daily security watch.

Built this CV’s AI avatar: an agent grounded on my CV and interview notes, whose answers cite the CV lines they draw on. A working demo of context engineering.

Value YOU, native iOS and Android companion apps for a workshop facilitator’s deck of appreciation cards, for a real client. Claude is the developer; I own the product and the delivery.

Mirror, an iOS app (SwiftUI, SwiftData, CloudKit) on Apple’s on-device Foundation Models, with a Cloudflare Worker proxy for the paid tier, hardened through Claude Code security passes.

Mjölnir, my personal dashboard on SOL, now being rewritten as a native macOS app by Claude Code from a plan I wrote.

Signal Drop, an AI podcast pipeline running on SOL.

Australian and New Zealand citizen, free to travel.

Roland reads every conversation.

My CV only tells you so much

My AI avatar knows how I work, what I’ve built and why. Ask it what you’d ask me in an interview. Its answers light up the lines of the CV they come from, and I read every conversation.