From France 🇫🇷, based in Vancouver, BC 🇨🇦 · English & French
13 years, 5 countries, one lesson: keep it simple.
These days I handle the engineering on UpMatch, a SaaS for freelancers. Before that I built products in teams of every size, from big platforms to three-person startups.
French, in Canada EN · FR ex-Amazon Alexa Sole engineer on a live SaaS
I started at the physical layer: three years as a network technician apprentice at Orange, wiring buildings and call centres for thousands of employees. Infrastructure was my first language.
Then software pulled me in. A developer diploma, internships building video platforms and medical-app APIs in Symfony, and a first real engineering job at Nelio, where I migrated every company service to Docker Swarm and shipped a React Native app used by 20+ shops.
02 · 2018-2019 · 🇳🇿 Auckland, New Zealand
CI/CD for 40+ developers, $8K/month shaved off the AWS bill.
DevOps Engineer · Roadpass Digital
On the other side of the planet, I ran deployments for the web properties of a holding that owned most of New Zealand's tourism attractions.
I built the Jenkins and TeamCity pipelines a 40-developer team shipped through, owned the PagerDuty on-call stack, and cut $8K/month of AWS spend through right-sizing, reserved instances, and ECS consolidation, without a single service regression.
-$8K/mo · cut from AWS spend with zero service regression
03 · 2019 · 🇰🇭 Phnom Penh, Cambodia
Millions of parcels a month, and revenue from $0 to $12K.
In Phnom Penh I led a three-developer team building Ship24, a package-tracking platform ingesting real-time events from 1,000+ courier services through AWS Firehose and Kinesis pipelines.
I designed the GraphQL API partners used to query nested tracking events, and I owned the business outcome too: monetization went from zero to $12K/month in eight months through page-speed work, SEO, and AdSense optimization.
Millions · of parcels tracked per month$0 → $12K · monthly revenue grown in 8 months on Ship24
04 · 2019-2021 · 🌏 Remote, Australia & France
20+ contracts across three continents. Rating: 5.0.
Freelance Solutions Architect · Upwork & direct clients
I went fully independent: TypeScript, React, and AWS serverless builds for SaaS clients across North America, Europe, and Australia. Text-to-speech systems, speech-to-text, automated call distribution.
To stop rebuilding the same foundations, I wrote an API design-system and scaffolding library reused across four client projects. It cut new-API delivery time by roughly 40%.
20+ · 5.0★ · freelance contracts delivered
05 · 2021-2023 · 🇫🇷 Paris, France
Voice experiences for three countries, and tooling that saved 100+ engineering days a year.
Solutions Architect, Alexa · Amazon
At Amazon I owned the integration, launch, and operation of consumer Alexa experiences across France, Italy, and Spain, with partners ranging from music streaming to radio, news, video, sports, and games.
The bottleneck was manual partner onboarding, so I built a TypeScript tool that automated Skill validation. It saved 100+ engineering days per year, was used weekly by 40+ Solutions Architects across the org, and cut two weeks from the Alexa Middle East launch.
100+ · engineering days saved per year at Amazon Alexa
06 · 2023-now · 🇨🇦 Vancouver, Canada
The only engineer on a SaaS 90 freelancers pay for monthly.
Founder-mode: solo products · UpMatch · StoreAI · cica.ai
Now in Vancouver, I build entire products as the only engineer: front-end, back-end, AWS infrastructure, and AI orchestration. As co-founder & CTO of cica.ai I shipped an AI compliance agent for construction on a schema-first AppSync GraphQL back-end.
Today I'm the sole engineer behind UpMatch, a multi-tenant SaaS that scores every new Upwork job in real time and auto-bids for 90 paying freelancers. The platform earns $40-50K MRR as a business — sales and support drive that with their own team, while I keep every part of the product running. I also build StoreAI, which turns a product URL into a finished Shopify store in under 60 seconds using Claude, ChatGPT, and Gemini.
Along the way: a Data Scientist master's (OpenClassrooms, 2024) and a published Edge Impulse case study on visual anomaly detection deployed to SageMaker Serverless.
$40-50K · monthly platform revenue~5 min · from job posted to AI-written bid submitted<60s · from product URL to a finished Shopify store
Today · Vancouver, Canada 🇨🇦
The route ends in Vancouver, for now.
I'm open to senior full-stack or solutions architect roles, and to contracts with real scope. Remote works, timezones aren't a problem, and I'm as happy owning production as I am pairing with a team.
Numbers from projects I've worked on, with the story behind each. Click any card for the full case study.
$40-50K▾monthly platform revenue, kept running by one engineer🇨🇦 UpMatch · 2023-now
Situation
UpMatch had a market and a sales team, but no engineering team: freelancers were losing jobs because they bid too slowly.
Task
Design, build, and operate a multi-tenant SaaS that bids on Upwork jobs automatically, end to end, as the only engineer (4 days/week).
Action
Built a real-time pipeline that polls Upwork every minute, scores each job against the freelancer's profile (skills, budget, location, client history), and triggers a headless-browser bidding bot with AI-generated proposals and Loom integration. Shipped the React dashboard, PostgreSQL multi-tenant data layer, and Terraform-managed AWS serverless infra with GitHub Actions CI/CD.
Result
90 paying freelancers generating $40-50K/month for the platform, with bids submitted ~5 minutes after a job is posted.
100+▾engineering days saved per year at Amazon Alexa🇫🇷 Amazon · 2021-2023
Situation
Alexa partner onboarding and Skill validation were manual, slow, and error-prone: a bottleneck for every launch across Europe.
Task
Own integration and launch of consumer voice experiences for France, Italy, and Spain, and unblock the pipeline.
Action
Built a TypeScript tool that automated Skill validation and partner onboarding; wrote the architecture docs, ran code reviews with partner engineers, and defined SLOs, alerts, and runbooks for skills serving consumer traffic.
Result
100+ engineering days saved per year, weekly use by 40+ Solutions Architects across the Alexa org, and two weeks cut from the Alexa Middle East launch.
$0 → $12K▾monthly revenue grown in 8 months on Ship24🇰🇭 Okoone · Ship24 · 2019
Situation
Ship24 had real traffic tracking parcels worldwide, and zero monetization.
Task
Lead the three-developer team, own the architecture, and make the product earn.
Action
Shipped the GraphQL API and real-time ingestion architecture, then treated performance as revenue: page-speed optimization, SEO, and AdSense tuning, iterating against analytics.
Result
$12K/month within eight months, while the platform tracked millions of parcels monthly.
-$8K/mo▾cut from AWS spend with zero service regression🇳🇿 Roadpass Digital · 2018-2019
Situation
The AWS bill for a national tourism holding's consumer properties kept climbing.
Task
Run infrastructure for 40+ developers and get spend under control.
Action
Right-sized EC2 fleets, planned reserved instances, consolidated workloads onto ECS, standardized deployments through Jenkins/TeamCity pipelines, and owned the PagerDuty + CloudWatch alerting stack.
Result
$8K/month saved, about $96K/year, with no service-level regression.
Millions▾of parcels tracked per month, from 1,000+ couriers🇰🇭 Okoone · Ship24 · 2019
Situation
A Hong Kong client needed unified package tracking across more than 1,000 courier services with wildly inconsistent data.
Task
Architect a platform that ingests and serves that volume in real time.
Action
Designed an AWS microservices architecture with Firehose + Kinesis ingestion pipelines and MongoDB storage, and a GraphQL API flexible enough for partners to query nested tracking events and carrier metadata.
Result
Millions of parcels processed per month, consumed by the Ship24 front-end and external partner integrations.
Independent, new market, no safety net: every delivery graded in public.
Task
Deliver full-stack products for SaaS clients across three continents, solo.
Action
Shipped TypeScript/React front-ends, AWS serverless back-ends, and GraphQL APIs (text-to-speech, speech-to-text, automated call distribution), and built an API scaffolding library reused across four projects, cutting new-API delivery time ~40%.
Result
20+ engagements completed with a 5.0 rating.
Selected work
Shipped products, each with its own deep-dive page.
Front-end, back-end, data model, infrastructure, CI/CD, monitoring: on the products I run today there's no one to hand off to, so I scope, build, deploy, and wear the pager.
That only works with deliberate tooling: serverless where it fits, Terraform for everything, boring tools chosen so a small team, or one person, can keep many products healthy.
02
Teams taught me most of it
The 40-developer platform group in New Zealand taught me operations at scale, Amazon taught me rigor and writing things down, and startup teams in Phnom Penh and Vancouver taught me focus. Working solo now is only possible because of what I picked up working with others.
03
Functional core, pragmatic edges
Immutable state, pure handlers, composition over inheritance. A solo-maintained codebase stays honest when the functions do one thing and the state can't mutate behind your back.
04
The business metric is the metric
Ship24 wasn't finished when the API looked good. It was finished when revenue went from $0 to $12K/month. Same for UpMatch: what matters isn't the architecture diagram, it's the 90 people paying every month. I optimize for the number the client actually cares about.
05
Five countries, two languages
France, New Zealand, Cambodia, Australia, Canada. Fully bilingual English/French. I've been the technical bridge between radio-station executives and Amazon engineers, and adapting to context is the most transferable skill I have.
