I design, build, test, and ship production-ready AI systems that transform manual workflows into intelligent business software. From AI-powered document intelligence to enterprise SaaS platforms, I engineer solutions that solve real operational problems.
Intelligent workflow automation powered by OpenAI. Document intelligence, AI-assisted review systems, and prompt-engineered pipelines that replace manual processes.
Production-grade multi-module platforms with complex backend logic, role-based access control, audit trails, and operational tooling built for real businesses at scale.
End-to-end business process automation. Commission engines, approval pipelines, lifecycle systems, and scheduled backend workflows that eliminate repetitive manual work.
Stripe, PayPal, OpenAI, Supabase, REST APIs, and webhooks. Full integration architecture from authentication and error handling to data mapping and reliability.
I'm an AI Product Engineer who designs, builds, tests, deploys, and maintains the software I create. End-to-end ownership — from architecture decisions to production reliability — is how I operate, not just how I describe myself.
I specialize in solving business problems using AI. That distinction matters. Every system I build starts with understanding what's broken operationally — where time is wasted, where errors occur, where humans are doing work that software should handle. Then I engineer the solution.
My approach: business-first automation, production-first engineering. I build reliable software systems that create measurable outcomes — not impressive demos that break in production.
Designing AI-native business systems that combine OpenAI, workflow automation, and enterprise SaaS architecture to replace manual operational processes with intelligent, reliable software.
Production systems across AI, SaaS, and enterprise
OpenAI, GPT-4 Vision, prompt engineering, AI pipelines
Canada · France · Germany
All maintained. All in production.
Every engagement starts with the problem, not the technology. What's the manual process? Where is time being wasted? What does a good outcome look like?
I integrate AI when it genuinely improves the outcome — not for show. Document intelligence, intelligent routing, prompt-engineered extraction, and AI-assisted decision making.
Business logic lives server-side. Approval chains, AI pipelines, commission engines, lifecycle transitions — all backend. Pages stay thin, fast, and maintainable.
Manual workflows get automated first. Then optimized. Query design, caching strategies, and pre-computed aggregations — applied deliberately, not by default.
Atomic state transitions, error handling, API fallbacks, audit trails. Reliability before complexity — systems that work every time, not just in demos.
I own what I build through production. Monitoring, iteration, debugging, and improvement — not just delivery. End-to-end engineering responsibility.
I use AI to solve real operational problems — document extraction, workflow intelligence, decision support. Not as a feature, as a solution to something that was genuinely painful before.
If a human does it manually more than twice, it should be automated. But automation without reliability is just faster failure — every workflow gets error handling, fallbacks, and audit trails.
Business logic never lives in the UI. AI pipelines, approval chains, calculations, and state transitions belong server-side — more reliable, more testable, easier to evolve.
Indexed query fields, paginated searches, pre-computed summaries. Scale is a design decision made at the start, not a problem solved after launch when it's already painful.
The goal is never "implement OpenAI." The goal is "reduce document review time by 80%." Technology is the means. Business outcomes are the measure.
I don't hand off and disappear. I deploy, monitor, debug, and iterate. If something breaks at 2am, I want to know. End-to-end ownership is how reliable software gets built.
OpenAI API, GPT-4.1 Vision, structured extraction, fraud signal detection, AI-assisted workflows, and production-grade prompt engineering.
Full-stack Bubble.io: backend workflows, recursive automation, multi-module SaaS, data architecture, performance optimization, and enterprise-grade applications.
End-to-end SaaS: multi-tenant architecture, subscription billing, admin tooling, RBAC, audit logging, and operational dashboards for real businesses.
API design, webhook architecture, error handling, retry logic, and third-party integrations that are reliable under real production load.
Stripe, PayPal, Supabase, Postmark, Unlayer, OpenAI, and sportsbook APIs. Full integration lifecycle: auth, data mapping, error states, and monitoring.
Relational data modeling, query optimization, indexed search design, pagination strategies, and data structures engineered for scale and reliability.
Architected and delivered a full enterprise device and inventory lifecycle management platform. Dual-architecture inventory system (bulk and serialized), device state machine, RBAC across four permission levels, automated boarding workflows, warranty monitoring, and immutable audit trails — designed in Figma and shipped to production.
I solve business problems using AI. If you have a manual process that should be automated, a workflow that needs intelligence, or a system that needs to be built properly — let's talk.