Full-stack developer · AI-first
30 years building and owning software — from the silicon to the cloud. Laravel and REST at the core, AI woven into how I ship and into the products themselves. I like owning a thing end to end and standing behind both the architecture and the craft.
I build the whole chain, not just one layer. I've written embedded firmware on my own sensors and the cloud platform they report into; I've owned high-volume billing where a wrong number is a real bill; and I structure my code so AI coding agents can work in it safely. I'm language-agnostic by temperament and ramp fast on new stacks — right now I'm deepening Node.js and Kubernetes the way I learn anything: by building something real. What stays constant is the discipline.
A Laravel Artisan command I built. It plugs into an N+1 detector — which I forked to support the
latest Laravel versions — takes the recorded N+1 query warnings, and uses the Anthropic Claude API
to generate the Eloquent fix and safely apply it to the code. I also structure my repositories
(CLAUDE.md / .claude)
so AI coding agents can operate in them safely — AI is part of my toolchain, not a bolt-on.
An end-to-end pipeline from physical sensors to the cloud. I built the whole chain myself: the hardware, the embedded firmware running on the sensors, and the Laravel / REST cloud integration they feed — running Docker (docker-compose + a web Dockerfile) with a Jenkins CI pipeline. It's my proof that I know both ends of an integration, not just the API contract.
I built this whole platform on my own for the customer Legtech — front end, back end and infrastructure. A Laravel REST API (JWT auth, activity logging, Excel export) backs a multi-tenant model of devices, jobs, alerts and threshold limits, with an Angular single-page app on top: live telemetry on Leaflet maps, ApexCharts dashboards, and an installable PWA. It runs on Docker with local parity to production, and I keep it current — most recently planning its PHP 8.5 / Laravel 13 upgrade. End-to-end ownership of a real product, for a real client, by one person.
When I decide to learn a stack, I build with it rather than read about it. This is a running Kubernetes cluster (kind) with two Node.js services and Postgres: a classifier flags sensitive content, and the message API AES-256-GCM-encrypts it at rest with a key held in a Kubernetes Secret. It has liveness/readiness probes, a Horizontal Pod Autoscaler proven under k6 load, TLS-terminating ingress, and a green CI pipeline that stands the whole thing up on a real cluster. I'm a 30-year engineer who's newer to Node.js in production — this is how I close that gap in the open. Code: github.com/jegjessing/zero-touch-encryption-lab.
A gas-billing module where correctness isn't optional — every invoice is a real bill to a real customer. High volume, high stakes, and the kind of domain where the tests matter as much as the feature.
I owned a multi-tenant SaaS compliance / GRC backend. Quality was a gate, not a hope: pre-push Git hooks enforced ~83% coverage on the business logic before anything could land. Discipline built into the workflow, so the whole team inherited it.