Open to AI & automation roles

AI products for messy expert workflows.

The hard part is not collecting more evidence. It is turning scattered evidence into a record people can trust.

View work → Talk through a workflow →
3+ Live AI products designed & built
10 yrs Built-environment delivery
UCL PhD-level research rigour
UK London · full right to work
01 Approach

Make uncertainty visible.

Before AI, I learned how trust is built in delivery: briefs, budgets, site evidence, handovers. That is the lens I bring to workflows where uncertainty has to be visible.

02 Selected work

Messy evidence in. Verifiable output out.

001 AI · Fintech
SlipMind ↗

Pre-accounting memory for sole traders

Live product · public GitHub demo

In
Receipts, invoices, forwarded emails, CSVs
Out
Reviewed records, evidence flags, accountant-ready exports
View live product →
002 AI · Land ops
GroundTruth PM ↗

Visibility & reconciliation for remote land assets

Working product system · private walkthrough

In
Satellite observations, WhatsApp updates, field photos, receipts
Out
Weekly visibility brief - claims reconciled, decisions needed
View working system →
003 AI · Construction
Bildeasy ↗

Compliance infrastructure for construction lending

Live product system · private walkthrough

In
Project files, site evidence, professional verification
Out
Audit-ready verification records for lending decisions
View product case study →
004 AI · PropTech
Ledger & Lease ↗

Private proof records for property compliance

Live product · payment-gated workflow

In
Emails, screenshots, PDFs, notices
Out
Dated proof records, verification links
View live product →
005 Delivery · Ops
DOS Design Studio ↗

Capital programme delivery in complex built environments

Client delivery work

In
Briefs, constraints, budgets, stakeholders
Out
Coordinated delivery, documentation, handovers
View case study →
03 Proof of work

Working products, not slideware.

Some work is public. Some stays private because the workflow is the value.

  • SlipMind
    Evidence extraction, structured records, accountant-ready exports.
  • GroundTruth PM
    Claim-evidence reconciliation, confidence scoring, decision briefs.
  • Bildeasy
    RAG-grounded compliance, human-in-the-loop verification records.
  • Ledger & Lease
    Proof records, verification links, payment-gated flow.
  • DOS Design Studio
    Capital delivery, stakeholder coordination, documentation and handover judgement.
04 How I work

From reality to record.

01

Map reality

Who acts, what arrives, where judgement happens.

02

Define the output

If the output is unclear, the model is decoration.

03

Design the AI layer

Extract, draft, flag uncertainty, hand off.

04

Ship the workflow

Working screens, records, dashboards and handover.

05

Test for failure

Missing evidence, false confidence, polished nonsense.

05 Principles
  • /01 AI should expose uncertainty, not hide it.
  • /02 The product is the record, not the conversation.
  • /03 Human review is part of the system, not a disclaimer.
  • /04 The best AI products feel boring once they work.
06 Stack

How these show up in real systems.

  • RAG so answers stay grounded.
  • Structured outputs that become records.
  • Human review gates where accountability matters.
Delivery / built environment
Capital programme delivery Stakeholder coordination Project documentation Site evidence Handover Risk and issues
AI engineering
LLM APIs Agent frameworks Structured outputs Tool calling RAG Document processing Workflow systems Prompt architecture
AI-assisted development
Claude / Claude Code Codex Cursor
Build stack (AI-assisted)
Next.js / React Supabase / Vercel FastAPI Stripe / Resend / Cloudflare Python (learning)
Design
Figma Photoshop / Illustrator / InDesign
Distribution & validation
Landing pages & waitlists Framer PostHog (familiar) Google Analytics (familiar) Meta ads (familiar) Reddit
07 Best fit

What to hire me for.

Best fit: work where decisions depend on records people can trust - evidence, compliance, field updates and operational handoffs.

Delivery PM - built environment Capital programme delivery AI Workflow Engineer Forward Deployed AI Operator AI Product Builder Automation / Product Ops AI Solutions Designer · evidence-heavy workflows Product-minded operator · regulated domains

Best where product judgement, operational context and engineering have to meet.

08 Contact

Bring me the workflow everyone works around.

For evidence-heavy, regulated or operationally messy work.

Private walkthroughs

Some commercial workflows and compliance logic are private. I share walkthroughs, architecture notes and redacted outputs on request.

EmailStart an email → LinkedInin/donald-oluoch → GitHubgithub.com/donaldoluoch →
LocationLondon, United Kingdom