Fox AI
A self-hosted control plane for cloud and local models, with one gateway, project-scoped access, spend attribution, and a private RTX inference path.
I work where engineering, AI, automation, and product design overlap — from production interfaces to self-hosted AI infrastructure and intelligence workflows.
A mix of production systems, internal tools, and experiments. The common thread is turning fragmented workflows into something explicit, observable, and useful.
A self-hosted control plane for cloud and local models, with one gateway, project-scoped access, spend attribution, and a private RTX inference path.
An intelligence workflow for grouping scam activity into campaigns, connecting infrastructure and behavioral signals, and turning fragmented evidence into operational leads.
A job-search automation system that discovers company career sources, normalizes ATS data, enriches listings, and pushes actionable workflows into Telegram.
A personal finance workflow for collecting transactions, tracking budgets and recurring costs, and turning messy household inputs into something actually usable.
I like work that crosses boundaries. A frontend problem is often really a product problem; an AI feature is often really an infrastructure and observability problem.
Frontend architecture, TypeScript-heavy applications, APIs, data flows, and the unglamorous production details that decide whether software survives contact with users.
Model gateways, local inference, agents, structured workflows, retrieval, cost controls, and automation designed around real operational constraints.
Taking vague operational pain, mapping the workflow, choosing the smallest useful system, and pushing it far enough to expose what actually matters.
The most interesting problems to me right now sit between software infrastructure, intelligence workflows, and practical AI adoption.
I’m interested in ambitious engineering, AI infrastructure, product systems, and intelligence workflows where the hard part is bigger than a single framework.