Engineering Services
Senior engineering judgment, in writing.
Whether you’re diligencing a target or making the case for your own platform, it’s the same discipline: architecture, data, and the deploy path, read by a fifteen-strong practice that has shipped in regulated environments, and written up so a board can act on it.
Run the free scorecardDisciplines
One team, the full bench.
Engineering starts before the first commit and continues long after the launch.
Where we go deepest Entity resolution and identity graphs in regulated data: financial crime, sanctions compliance, and research platforms.
Staffed across the board A fifteen-strong practice spanning cloud, data, security, and delivery, so the recommendation and the build don’t have to come from different firms.
Architecture & platform
Architecture and systems design
Systems drawn to survive their second year, not just their launch. The trade-offs made in advance so they do not surface as outages later.
Cloud infrastructure
AWS, Azure, GCP, and the pragmatic mix. Infrastructure that reflects the actual load rather than the aspirational one, sized for cost as much as capacity.
Legacy modernisation
Strangler-fig migrations, phased cutovers, zero-downtime rewrites. Moving off the old thing without stopping the business it runs.
Data, AI & security
Data engineering and governance
ETL, pipelines, warehouse design, data lineage, PII handling, audit trails. Kept clean enough for your auditor as well as your product team.
AI integration with guardrails
Retrieval, agents, and model orchestration deployable inside regulated environments. PII stays out of vendor models. Audit trail included.
Security and compliance engineering
SOC 2, HIPAA, ISO 27001, FedRAMP shaped into your SDLC. Threat modelling, access review, secrets rotation, pen-test remediation.
Assurance & leadership
Technical due diligence
Pre-deal reads of the platform, team, and failure modes. Risk maps, cost drivers, and quantified remediation plans that stand up to board scrutiny.
Fractional engineering leadership
A senior engineering seat inside your weekly cadence: technical strategy, delivery discipline, vendor management, and the calls only a real engineer should make.
What we can package
Most engineering work is bespoke. Some of it we can just book.
Most of the value we add lands as bespoke work. But a few of the reads and deliverables scope cleanly enough to book directly, at prices you can see up front. Start with the free scorecard to find out whether we are the right team for what comes after.
Fixed-fee deliverables
Book any of these directly. Priced up front.
Audit
Architecture Audit
A senior look at the system as it is, modelled against the change you are planning. What breaks, what holds, and the cleanest path forward.
See the architecture auditRoadmap
Modernisation Roadmap
Phased plan to move off a legacy system without halting the business. Milestones, risk assessments, and the handover points where your team takes over.
Discuss modernisation roadmapDeep read
Technical Assessment
A senior read of your platform. Architecture, code, data, and the deploy path, interviewed with the team that owns them. Board-ready written report.
Discuss technical assessmentFree, on us
Tech Risk Scorecard
The interactive version of the technical reads we run for money. Twelve questions, six dimensions, a tier score. Runs in your browser, in about twenty minutes.
You get a tier score for your platform and the flags that matter most. If you want a senior second opinion on the result, we write one back within two working days. If you do not, your answers stay on your machine.
Interim engineering leadership
When you need engineering leadership in the room, not just on the org chart.
A senior engineering leader embedded inside your weekly cadence, holding the seat that keeps the roadmap, the architecture, and the release calendar moving in the same direction. Three-month minimum, fixed monthly fee, handover documented from day one. The engagement is designed to end.
Who shows up Tom Barber: NASA/JPL (Mars 2020 PIXL instrument), Princeton (20-person data engineering team), VP at the Apache Software Foundation, and financial-crime compliance platforms.
And behind the seat A fifteen-strong bench. When the fractional leader finds the problem, the practice can fix it across architecture, data, security, or delivery, without a second procurement cycle.
The playbook we run The seat comes with a method, not improvisation: the First 90 Days playbook, the week-by-week plan we work through from the first day in the seat.
Who does the work
Named leads. A bench you actually meet.
Led by
Tom Barber
Senior Solutions Architect. NASA/JPL (Mars 2020 PIXL instrument), Princeton (20-person data engineering team), VP at the Apache Software Foundation, and financial-crime compliance platforms.
Engagement leads
- Juan Resendiz, Engineering Manager, ex Altech USA
- Marcin Ponichtera, Cloud Engineering Lead, K2 Integrity
The bench
Fifteen in total across cloud, data engineering, security, and delivery. Some are embedded with clients under NDA, so not everyone is on this page, but you meet the engineers who will staff your engagement before you sign, not after.
AI in our practice
Fluent across the stack. Not tied to any one tool.
AI is in our engineering loop every day. Claude, Cursor, Codex, agents, coding assistants, models we run locally, models we call from a shell. We pick the one that fits the loop, not the one on a vendor slide, and we keep up as the stack shifts.
Where AI touches a compliance surface, PII, production, or a security-critical path, we run it with guardrails. Model choice honours data residency. No client PII enters vendor models. Sensitive changes get human review before merge. Local and on-prem model options are on the table when the regime requires them. That is how we can put AI into regulated environments where most teams still cannot.
What we build with AI is portable, and every AI-touched change is traceable in your commit history. Custom Claude skills, agent loops in your CI, testing harnesses shaped around your patterns. Your team owns the loop, and your audit team can trace what a model wrote back to the prompt that produced it.
Start small
Run the tech risk scorecard.
About twenty minutes, free, in your browser. You get a tier score for your platform and the flags that actually matter. We write back a senior second opinion if you want one.