← Back to home /ABOUT

SKILLS

I design, build and run web applications — from the first mockup to the server that delivers them. What interests me is not technology for its own sake, but what it settles: a process that drags, a file nobody dares open any more, a fraud discovered too late, a site that goes down on a Saturday night.

Working across the whole chain alone has one concrete advantage: the person writing the code is also the one who configured the server and who will read the logs when something goes wrong. Nothing gets lost between two teams.

01Websites and web applications

From a showcase site to a complete business application, in PHP and JavaScript, with no dependency on an imposed framework. Every project is built to last and to be picked up again — by me or by someone else.

• Custom development — no purchased template bent out of shape to fit: the structure follows the need, not the other way round.
• Responsive interfaces — usable by thumb on a phone and by keyboard on a wide screen, tested at both ends.
• Accessibility and bilingualism — verified contrast, keyboard navigation, content served in French and English under distinct addresses.
• Performance — fonts and assets hosted locally, a single stylesheet request, no third-party script to slow you down or watch your visitors.
• Data integration — API consumption, official data feeds, dashboards and charts fed in real time.

02Business tools: the end of the shared spreadsheet

Many companies run part of their operations on a workbook nobody dares change: nested formulas, macros written by someone who has left, diverging copies on three machines, and a single person able to touch it. It works — until the day it doesn't.

I replace those files with real, small web applications, reachable from a browser, with nothing to install:

• A single source of truth — no more “v3_final_REVISED.xlsx” making the rounds by email.
• Concurrent access — several people work at once, each with their own permissions.
• Controlled input — fields validate what goes into them instead of letting a typo silently skew a calculation.
• History — who changed what, and when. A workbook will never tell you.
• Automation — recurring reports, reminders and exports generate themselves, on schedule.
• Carrying the existing over — the workbook's data and logic are recovered, not discarded: the transition loses nothing.

03Fraud prevention

Detecting fraud starts with knowing what normal looks like — then building the controls that make deviations stand out, without drowning teams in false alerts.

• Behaviour analysis — spotting abnormal patterns in transactions, sign-ups or usage.
• Rules and thresholds — automated controls, tuned to limit the false positives that wear vigilance down.
• Defence in depth on forms — anti-forgery token, bot trap, minimum delay, rate limiting, content filters: several layers, none of which is a single point of failure.
• Traceability — logs designed for investigation, holding what it takes to reconstruct an incident and nothing more.
• Investigation — cross-referencing sources, rebuilding timelines, documentation that holds up in a dispute.

04Cybersecurity

Security is not a layer added at the end. It is decided when the structure is chosen, and verified continuously thereafter.

• Application hardening — strict content security policy, complete security headers, cookies flagged Secure, HttpOnly and SameSite.
• Injection prevention — prepared statements, systematic output escaping, input validation at the boundary.
• Encryption in transit — HTTPS only, HSTS, automatically renewed certificates.
• Minimisation — not collecting what you don't need remains the most reliable protection against a leak.
• Québec Law 25 compliance — a privacy policy grounded in what the code actually does, assessment of transfers outside Québec, retention periods actually honoured.
• Code and infrastructure review — reading an existing system to identify what exposes it, with fixes prioritised by real risk.

05Infrastructure, hosting and operations

A site only truly exists once it is served, monitored and backed up. I administer the whole chain, from the operating system to the domain name.

• Linux servers — installation, hardening and maintenance on Ubuntu and Debian.
• Nginx and PHP-FPM — fine-grained configuration, rewrites, caching, performance tuning.
• Databases — MySQL and MariaDB: schema, indexes, verified backups — a backup never restored is not a backup.
• Automation — Bash scripts, scheduled tasks, reproducible deployments.
• Monitoring — structured logs, alerts on outages and intrusion attempts, controlled rotation and retention.
• DNS, certificates and CDN — zones, email, automated TLS, upstream proxying and protection.

06How I work

• Understand before coding — the right tool comes from the right question, not from a copied specification.
• Deliver in steps — something usable early, rather than a large project discovered at the end.
• Verify rather than hope — what ships has been measured, not merely re-read.
• Document — the code explains the how; comments and documentation explain the why.
• No forced dependency — you must be able to carry the project on, with me or without me.

Technical stack