Technical Blog

4 Min Read

A checklist before pointing GitHub Copilot's modernization agent at a legacy codebase: test coverage as a prerequisite, review capacity as the real constraint, and what to keep humans in front of.

4 Min Read

Pipeline duration is an observability problem, not a performance problem. Why CI runs deserve traces rather than transcripts, what usually turns up, and why flakiness costs more than slowness.

5 Min Read

Taking LLM-assisted ICD-10 coding from benchmarking to production — evaluation on real clinical data, guardrails, human-in-the-loop design, and honest notes on GPU capacity planning.

4 Min Read

Why integration tests against real PostgreSQL containers and metric-gated canary rollouts solve different halves of the same problem — and why you should adopt them in that order.

5 Min Read

Eighteen months into an engineering manager role: the leverage shift, hiring leads, why multi-cloud is an organisational problem, and the two metrics I actually watch.

4 Min Read

A service template is how an engineering org makes its standards the default instead of the aspiration — and why drift, not creation, is the problem that kills most of them.

4 Min Read

A decision framework for real-time updates in .NET: why server-to-client problems usually want Server-Sent Events, what bites you at scale, and an honest defence of polling.

4 Min Read

dotnet run is restore, build, and a runtime bootstrap wearing a trench coat. Why startup time became a scaling property, where it actually goes, and what genuinely helps.

4 Min Read

The RequestDelegate error in Minimal APIs as a case study in diagnostics: why overload resolution reports against the wrong overload, how to narrow it in ninety seconds, and the general technique.

5 Min Read

Three production PostgreSQL incidents after a large migration, and what each one teaches about indexing strategy, HOT updates, fillfactor, and why query plans change without a deploy.

4 Min Read

Editor tooling doesn't set your standard — the repository does. On enforcing analyzers in CI, keeping a noise budget, separating style from quality, and why this matters more in AI-first editors.

4 Min Read

Better Git tooling won't give you short-lived branches or reviewable pull requests. On why branch lifetime is the only branching question that matters, and why large PRs are effectively unreviewed.

4 Min Read

AutoMapper and MediatR going commercial is a useful case study in dependency risk: separate cost from coupling, do the usage inventory, then choose among paying, pinning, replacing, or removing the need.

4 Min Read

A year of AI-assisted code review: the mechanical defects it reliably finds, the domain correctness it structurally can't, and the real risk — what automated comments do to reviewer attention.

4 Min Read

A source generator is a build-time dependency, and build time is a shared resource. What generators actually cost, the packaging trap, the debugging question, and when they're clearly worth it.

4 Min Read

An honest account of what got faster, what didn't, and the new failure mode worth watching — confidence outrunning understanding, in a domain where 'looked right' isn't a standard.

5 Min Read

Lessons from taking a claims platform through NPHIES vendor certification — profile validation, identifier mapping, terminology, asynchronous messaging, and why the error paths carry the weight.

5 Min Read

How we moved a live claims platform off Azure CosmosDB and Elasticsearch onto PostgreSQL as the single source of truth — 50M+ records at terabyte scale, with dual-writes, shadow reads, and 100% reconciliation.

4 Min Read

IMeterFactory is the easy part. Choosing metrics is the hard part: rate, errors, duration, saturation — plus the cardinality trap and why business throughput catches failures system metrics never will.

4 Min Read

Your property names are public API. On separating wire types from domain types, deciding what unknown fields mean, explicit discriminators, and why nullable annotations lie at the boundary.

6 Min Read

Practical advice for software engineers transitioning to leadership roles, based on my journey from Senior Engineer to Technology Director.

5 Min Read

Real-world insights on implementing AI and ML in healthcare systems, from intelligent claims processing to clinical decision support.

5 Min Read

A practical guide to multi-cloud architecture based on real experience deploying healthcare systems across Azure, AWS, G42 Cloud, Oracle Cloud, and GCP.

5 Min Read

A candid reflection on founding Core System, the challenges of running a startup while working full-time, and why I made the difficult decision to close it.

5 Min Read

Practical insights on designing and scaling healthcare IT systems including EMRs, HIS, and claims management platforms.

5 Min Read

A personal reflection on my career path from IT support to leading technology strategy at a healthcare company, and the lessons learned along the way.