Builds you can
take apart.
No mockups, no mood boards, no "selected work coming soon". These are systems that run today, described at the level of engineering they were built to. Ask us anything about any of them - the answers are the interview.
Lumen v2 - a clinical assessment platform in production
A psychological assessment platform that administers a strengths-first instrument and produces separate participant and clinician reports from deterministic scoring. Not a prototype, not an MVP - it runs, it scores, it reports.
The engineering underneath
- ▶ Strict-TypeScript monorepo with Zod contracts end to end
- ▶ Supabase/PostgreSQL with row-level security across 35+ tables
- ▶ Deterministic scoring engine - same inputs, same outputs, always
- ▶ Regression-locked 301-field output contract with test coverage
- ▶ Separate participant and clinician report projections
Why this matters to your project
This is the standard of build behind every Diaboliq system engagement. When we say "proper databases, authentication, permissions", this is the reference implementation. The patterns are proven, so your build starts on solid ground instead of experiments.
National organisation website + knowledge dashboard
A production public website for an Australian national organisation, plus an internal knowledge-systems dashboard navigating a 140-page operational wiki. Both built for longevity: zero-dependency static pipeline, no CMS licences, no maintenance treadmill.
What was built
- ▶ Public-facing national website, mobile-first, fast on ordinary connections
- ▶ Zero-dependency Node build script generating the site data layer
- ▶ Dashboard with six views over the wiki: overview, browse, search, graph, activity
- ▶ Vanilla JS frontend - no framework tax, nothing to break yearly
Why this matters to your project
Total cost of ownership is a design decision, not an accident. Sites built like this have no licence fees, no plugin vulnerabilities, no annual rebuild cycle - the opposite of the WordPress treadmill most businesses are stuck on.
AI operations stack - agents doing daily work
Automation infrastructure running daily across communications, scheduling, knowledge retrieval, and content operations. n8n pipelines, AI agents with real tool access, and Model Context Protocol integrations - the same capability applied to client automation sprints.
What runs today
- ▶ n8n workflow pipelines handling recurring operational tasks
- ▶ AI agents with governed tool access - they act, within limits
- ▶ MCP integrations connecting agents to real systems safely
- ▶ Human-in-the-loop checkpoints wherever judgment matters
Why this matters to your project
Most "AI agencies" sell demos. This stack does weekly work. An automation sprint with Diaboliq means wiring proven patterns into your business - measured in hours saved, not slides presented.
Want this standard on your project?
Services and pricing are fixed and published. The intake form takes five minutes.