Workday integration consultant building AI-powered workflow tools
I've spent 5+ years implementing enterprise software — the last three leading Workday integration projects at a global consultancy, connecting HR and finance systems that companies run their business on. Now I'm applying that delivery discipline to LLM-based systems: agents, retrieval pipelines, and automation that hold up outside a demo.
As integrations lead I own the full lifecycle — requirements, solution architecture where standard templates don't fit, vendor coordination, migrations, and go-live — acting as the connective tissue across workstreams that turns business requirements into integration design.
Led the integrations workstream — designed and built the integrations tracking member retirement contributions and remitting pay actuals to the county payroll system. The firm's first public-sector engagement, opening the door to further government implementations.
Led integrations for a time-tracking implementation — connecting Salesforce time entry into Workday and reporting time and project data downstream to PeopleSoft.
Leading integration delivery — designed the inbound integrations bringing new-hire and time/project data from PeopleSoft into Workday, handing off the build to the client's MuleSoft team.
Leading integration delivery on a first global implementation. Hands-on with the client's business-technology and security teams connecting Boomi, ConductorOne, and Jira Service Management to Workday, with benefits and global payroll integrations in scope.
A retrieval-augmented assistant for Workday integration questions — hybrid vector + keyword search over a curated corpus, grounded answers with citations, and an eval suite measuring retrieval hit rate and answer faithfulness.
An LLM agent that turns unstructured contractor job requests into ready-to-submit permit packets — multi-step tool use over live jurisdiction data and Airtable, with a human approving every submission.
A risk-gated automated trading system: deterministic strategy engine, hard position-sizing limits, kill switch, and a full JSONL audit trail of every decision. 50+ tests on the risk and execution paths.
Full case studies — with outcomes, architecture notes, and demos — are coming as these ship.
A browser-based XSLT 3.0 editor with live output — plus a Claude agent that writes the stylesheet in front of you, runs every draft, reads the compiler errors, and fixes them. Bring your own API key; XML never leaves the browser except to Anthropic when you ask.
Upload a sample CSV audit log and get back a self-contained HTML report template with substitution tokens matched to your columns — drop it into a Workday Orchestrate flow and every run produces a filterable, searchable audit report.