AI-native systems builder & technical operator
I turn messy operational workflows into reliable software — across infrastructure, APIs, internal tools, data platforms, automation, and AI.
Selected systems
Each case study starts with the operational problem — then shows how the system was made reliable enough to own.
AI-powered email infrastructure observability
A production operations layer that helps mail operators understand why delivery is failing — before the incident becomes a support queue.
Correlates SMTP telemetry, DNS changes, DMARC data, reputation signals, and incidents. Deterministic evidence comes first; AI assists with diagnosis.
AI-native business operations platform
A sanitized demonstration of production architecture for turning fragmented commercial workflows into one accountable operating layer.
Connects research, qualification, AI drafting, human review, campaigns, client workflows, payments, reporting, and role-based operational visibility.
AI video production engine
A system that turns a topic or script into a finished video through an AI editorial layer and a deterministic rendering pipeline.
AI plans. Deterministic systems execute. Structured edit plans, workers, provider fallbacks, asset reuse, cost controls, and recovery make the pipeline operable.
How I work
The correct action should be the easiest action.
Map the actual business process before writing code.
Define authoritative data, workflows, failure states, and human responsibilities.
Use the simplest appropriate tool: application code, automation, AI, or infrastructure.
Deploy, monitor, recover, document, and improve.
Architecture philosophy
Good systems clarify what happens next — including when something fails.
01Deterministic core, probabilistic edge.
02Business state belongs in proper software, not prompt memory.
03Use AI where reasoning adds value.
04Use automation where the workflow is deterministic.
05Humans approve consequential decisions.
06Make failures visible and recoverable.
07Keep systems boring where boring is safer.
08Do not add another SaaS tool unless it solves a real problem.
Foundation
I’ve spent 17+ years building and operating web and infrastructure systems: hosting, Linux, cPanel / WHM / WHMCS, migrations, and hundreds of production environments. That operational history shapes how I build AI-native platforms now.
View experienceTechnical range
I work across the seams where most internal systems become difficult.
Linux, Docker, networking, hosting, monitoring, deployment
Python, FastAPI, Go, Node.js, TypeScript
PostgreSQL, Redis, ClickHouse, pgvector, SQL
LLM APIs, tool calling, RAG, structured outputs, agents, human-in-the-loop systems
Queues, retries, idempotency, auditability, observability, failure recovery
REST APIs, webhooks, Gmail, Google Workspace, payments, CRM, third-party SaaS