Work

Recent work, checked against real data.

Five recent engagements and one set of demos, with outcomes only where they were measured. Clients of the consultancy I work with are described, not named; if you hire me, the contract is with me directly.

Six entries, newest first. Select one to jump to it.
International human rights NGO · via Relief Applications219/222

A payroll allocation platform, built to replace Excel.

Context

The client’s staff allocated payroll across programmes in Excel workbooks. I built the timesheet and payroll allocation app that replaced them, which staff now use to close each month.

What I did

  • Built it: Flask, PostgreSQL and React.
  • Checked it against the client’s workbooks; found and fixed 17 defects.
  • Fixed controls that reported “compliant” on empty months.
  • Guarded closed months against later imports.
  • Five fixes shipped as five pull requests in one day.
More
  • Shipped staff requests: split allocations, Excel paste, multi-month uploads.
  • Unblocked a month-close traced across six layers of the app.
  • Wrote client updates in French; handed over once the client ran it alone.

Outcome

Coverage totals matched to the cent, 219 of 222; the only material gap was the client’s own workbook error. 2025 payroll: 87% of employees to the cent. Eight people paid but missing from the roster, flagged.

Python / FlaskPostgreSQLReactpytest

The full numbers →

Humanitarian accreditation body · via Relief Applications8 sections

An AI usage policy built around the data.

Context

Consultants and auditors needed clear rules on what could go into AI tools, including personal accounts.

What I did

  • Drafted the policy, tied to what data leaves the account.
  • Three kinds of account; a green, amber and red table.
  • Eight leadership decisions, each with a suggested answer.
  • A staff awareness programme and a one-page guide.
More
  • Reviewed the IT documentation and phishing-response guidance.
  • Compared options for connecting Claude to SharePoint and NAS stores.

Outcome

A draft with the client’s leadership for sign-off.

AI usage policyData protectionStaff awareness

What a policy covers →

International institutions · via Relief Applications4 deliverables

AI readiness assessments for non-technical leadership.

Context

Institutions in conservation, development and disaster-risk financing, asking where AI fits.

What I did

  • Current-state audit.
  • Opportunity mapping.
  • Vendor comparison.
  • Recommendations for leadership.

Outcome

Written recommendations for leadership. Supported, remotely, an in-person awareness programme in London.

Readiness assessmentVendor comparisonStaff awareness

AI advisory →

Ten Lives · regenerative community festival, PortugalPipeline

Supplier data, from scattered to searchable.

Context

The supplier information the festival needed was spread across Portuguese directories.

What I did

  • A scraping and enrichment pipeline across those directories.
  • Deduplicated; filterable by region, certification and category.
  • A volunteer onboarding pipeline, still in progress.

Outcome

One deduplicated, filterable supplier list. The volunteer pipeline has no outcome yet.

PythonGoogle Places API

Automation →

Rapidos.ai · Growth & Automation Advisor, contractn8n

Lead qualification and knowledge bases in n8n.

Context

An AI studio, August 2025 to January 2026.

What I did

  • Designed automated knowledge bases and lead-qualification flows.
  • Turned business requirements into n8n workflows and internal AI agents.

Outcome

n8n workflows and internal AI agents. No metric was tracked.

n8nn8n AI agents
My own work · sample data, not client work7 demos

Seven operational dashboards, on synthetic data.

Context

Client work can’t be shown, so these show the shape of a dashboard engagement instead.

What I did

  • Each reads a visitor’s own CSV or XLSX in the browser.
  • A Content-Security-Policy that stops them sending data anywhere.
  • Published security, deployment and GDPR documentation.

Outcome

Open to anyone: drop in your own file. It is read in your browser and sent nowhere.

JavaScriptChart.jsPython generators

See the dashboards →

Something like this on your side?

A 20-minute call about one problem, free. If it isn’t something I should do, I’ll say so.