Services
Most engagements combine two or three of these. Each is described by the problem it solves, what you get, and what it is built with.
Operational software
For a business that takes bookings, sells at a counter, schedules people and resources, or needs staff and customers to see the same information. Built as a web application that works on the phone in a staff member's pocket and the terminal at the desk.
What you get
- A written specification you approve before work starts, with the decisions you made recorded in it
- Working software every week, deployed where you can use it
- Automated tests for the parts that handle money and bookings, and a live check of every fix on the real system
- Source code in your repository, documented so another engineer can pick it up
Built with
Java and Spring Boot on the server, server-rendered pages with plain JavaScript, PostgreSQL or MongoDB, nginx. Boring on purpose: these have stable upgrade paths and will still be maintainable in 2031.
Payments, hardware and third-party services
The parts that sit between your software and the outside world, and the parts most likely to stall a project.
- Card payments and refunds through Stripe or Authorize.net, including tap-to-pay terminals
- RFID wristbands for check-in and spending; smart lockers through the manufacturer's cloud API
- Google sign-in, Gmail and Calendar access with encrypted credential storage designed to pass Google's app verification
- Transactional email, SMS through Twilio, and webhooks in both directions
Every integration is wrapped behind an interface with a recorded-response test, so when a vendor changes its API the break is found by a failing test, not by a customer at the counter.
Pipelines and published statistics
Pulling data from state agencies, partner websites or your own systems on a schedule, computing the figures people actually want, and publishing them on pages that search engines rank.
- Polite crawlers and feed ingesters: rate-limited, honest user agents, robots.txt respected, fixture-tested so a redesigned source is caught the day it changes
- Server-side computation of odds, expected values, depreciation curves and similar, so the method stays yours
- Freshness monitoring: a page serving stale numbers is treated as an outage and raises an alarm
- Technical SEO as part of the build: canonical hosts, sitemaps, structured data, crawl-budget hygiene
Hosting and operations
Your system on a server I manage, or on yours with me on call. Either way the goal is the same: you do not think about it.
- TLS certificates, nginx, systemd services that restart themselves, JVM guardrails so a memory fault becomes a clean restart rather than a silent hang
- Nightly backups with retention and a verified restore
- Health endpoints and daily alarms for failed jobs, stale data and failing integrations
- Incident response with a written cause, not a reboot and a shrug; infrastructure moves (one recent client moved from Google Cloud to Azure) planned and executed
How an engagement works
- Scope A short conversation, then a written design covering what will be built, what will not, and the decisions that are yours to make. You approve it before any code exists. Usually one to two weeks.
- Build Weekly deliveries of working software to a place you can use it, with tests for anything that touches money or bookings. Scope changes are written down the same way the original was.
- Run Launch, then either managed hosting with alarms and backups, or a documented hand-over to your team. Fixed monthly fee for hosting; time and materials for new work.