Currently

What I'm working on now

A snapshot of what's taking my attention this quarter. Updated when it changes.

Last updated: May 2026.

Day job

I'm Head of Architecture at Sonnedix, a global renewable-energy independent power producer. The work for the next two quarters splits across three streams.

Rolling out Meridian across the senior leadership team and their direct lieutenants. The application portfolio module is populated; the CANVAS approval workflow is folded in; the conversational assistant is in private beta with a handful of executives. The work for the next two quarters is the broadening: more stakeholders onboarded, more questions answered against the corpus, more integrations into adjacent systems.

Extending CANVAS with the post-contract workflow steps — CMDB onboarding, operational handover, the full serviceability checklist. Steps eleven through nineteen of the lifecycle that the MVP deliberately omitted.

Standing up architecture review for acquisitions as a repeatable practice. The framework is roughly the one I wrote up recently — 30-60-90-180, written assessment, categorised decisions, vendor consolidation calendar. Operationalising it across the function.

Writing

I'm publishing one long-form piece a week through 2026, alternating between case studies and commentary. The topic backlog at the moment includes: a deep technical piece on the data residency patterns inside Meridian, a follow-up to the audit-framework piece covering eval design specifically, and a piece on the role of the Chief Architect as the function evolves.

The full chronological list is at /blog/. The Atom feed is the right way to subscribe.

Reading

The pile in active rotation:

  • Continuous Architecture in Practice, Erder, Pureur and Woods — rereading after several years; the framing still holds up well.
  • Working in Public, Nadia Eghbal — the sustainability argument for open infrastructure, applied with adjustments to the enterprise context.
  • Recent papers on agentic LLM evaluation — particularly the work on benchmark contamination and the design of held-out evaluation sets that survive contact with frontier models.

Side projects

Springboard continues to evolve in small increments. The current focus is on the eval framework around the LLM-based application-fit scoring — making the scores more defensible and the failure modes more visible.

Open conversations

I'm open to:

  • Architecture advisory conversations with executive teams of fast-growing, acquisition-heavy companies, particularly those wrestling with the build-versus-buy question on EA tooling.
  • Speaking opportunities on AI in regulated enterprise, enterprise architecture practice, or large-scale transformation. See /speaking/ for past talks and topics on offer.
  • Selective senior architecture leadership opportunities (Head of Architecture, Chief Architect, Head of AI Architecture).

The right place to start a conversation is LinkedIn or the contact form on the homepage.


This page is in the /now/ tradition established by Derek Sivers — a single, deliberately current snapshot of what's taking my attention. The home and CV pages cover where I have been and where I am positioned; this page covers what is in front of me right now.