Applied AI Engineer and software engineer shipping end-to-end products — now focused on the systems around AI: workflows, APIs, voice, security, billing, observability, and deployment.
Four different AI product problems. Each case shows the surrounding system: product decisions, integration boundaries, failure modes, and evidence — not a list of model names.
A multi-tenant product that lets shoppers talk through high-consideration purchases while the interface responds with live product options and an order flow.
A monitoring system that turns website changes into prioritized business intelligence through scheduled collection, deterministic diffing, structured LLM analysis, and reports.
A scanner that inspects how AI chat products are configured in the browser, runs targeted probes, grades findings, and produces remediation-oriented evidence.
An end-to-end SaaS surface for handling clinic calls, appointment intent, urgent cases, transcripts, billing, and operator workflows.
Models create capability. The surrounding system determines whether that capability becomes dependable software, a usable workflow, and a business result.
Bounded model tasks with explicit inputs, outputs, and review paths.
Durable state and boundaries around probabilistic behaviour.
Interfaces where users can inspect, operate, and correct the system.
The controls that keep a product useful after the first successful demo.
Mobile releases, BLE protocols, offline-first products, commerce, and client delivery are supporting evidence: the engineering depth existed before the current AI cycle.
I am most useful where a capable prototype has to become a product that other people can operate and trust. I work across architecture, backend, product UI, integrations, deployment, and the uncomfortable details between them.
My earlier work spans native mobile, BLE hardware, health products, SaaS, and e‑commerce. That range now informs how I build applied AI systems: with explicit boundaries, real operator workflows, and respect for production failure modes.
Show the system, its status, and what can actually be verified.
Use conventional software to contain probabilistic behaviour.
Logs, controls, reports, and review flows are part of the product.
Deployment and maintainability matter as much as the first demo.
Short technical positions from current project work. These will connect to deeper writing as the new portfolio becomes the evidence layer behind LinkedIn.
What has to exist around an AI workflow before real users can depend on it.
Why conversation alone is not enough for commerce, booking, or support.
Configuration evidence, exposed data, and unsafe rendering before exotic attacks.
I am looking for Applied AI Engineer, Software Engineer, AI Product Engineer, backend, or product engineering roles with global remote teams, including US and Canadian companies hiring internationally.