Experience
An independent practice, taken from problem statement to deployed system and handover — alongside backend and machine learning work inside product teams. Everything on this page is verifiable; there are no client names where a client asked for none, and no metrics that were never measured.
Chronology
2024 — Present
Independent practiceCurrent
Direct client engagements
Business software, AI systems, frontend products and automation built for founders, small businesses and agencies — scoped, built and handed over as working software rather than as proofs of concept.
The independent practice does not pause between the roles below, which is why two of these periods read Present. Client engagements are scoped to fit around whatever the current role is, and taken only when they can actually be finished — that is the accurate version, and it is more use to you than a timeline implying several full-time jobs at once.
Project durations on the case studies are build effort, not elapsed calendar time. Add them up as calendar spans and they exceed the years they sit in, because the work ran alongside study and the roles below rather than instead of them — several engagements were part-time, overlapping, or picked up and put down around a term or a notice period. The effort figure is the honest one, and it is the one that tells you what a build like that would cost you.
Professional experience
Full-timeCurrent
AI features inside production software, across prototyping, model integration and the deployment path.
Internship · remote, Canada
Data analysis and machine learning on live project work, coordinated asynchronously across time zones.
Internship
Django REST layer and PostgreSQL schema behind a consumer community product, deployed as containers on Cloud Run.
Internship
Classification models on real datasets, with the evaluation treated as part of the deliverable and the results reported to non-technical stakeholders.
How I work
Stated up front because they are the reason to hire an engineer rather than a vendor, and because a client who disagrees with them should find that out early.
The first conversation is about what breaks today, who it breaks for and what a fixed version looks like. Technology choices come after that, and they are usually boring on purpose.
Most business software fails at the schema, not the screen. Getting the entities, states and audit requirements right early is what makes the second month cheaper than the first.
Anything a model decides gets a confidence, a source and a reviewable trail. A system nobody trusts gets switched off, regardless of how accurate it is.
Something usable exists early and grows. That keeps feedback cheap and stops a build drifting away from what the business actually needed.
Architecture notes, environment setup, data model documentation and a walkthrough. The work should survive without me being on call for it.
Engagement models
Scope and duration are indicative, not a price list. A written proposal with a fixed scope and timeline comes before any commitment.
Typically 6–12 weeks
A business system built from problem statement to deployed software: data model, interface, permissions, integrations and handover documentation.
Founders and operators replacing a spreadsheet-and-email process with real software.
Includes
Typically 4–8 weeks
Adding extraction, scoring, retrieval or conversational capability to software that already exists, with the review and audit path built in from the start.
Teams with a working product and a process that is still manual inside it.
Includes
Per project or retained
Engineering capacity behind an agency or consultancy, working to your standards and communicating through your process — white-labelled where required.
Studios that have sold the work and need someone to build the difficult part.
Includes
Education
Recognition
This list is short by choice. Vendor course badges and self-issued certificates are not listed, because a technical reviewer discounts them and an honest page is worth more than a padded one.
Availability
Remote engagements with founders, product teams and agencies. Send the problem, the constraints and the deadline — you will get a straight answer on fit within a working day.