Skip to content
OZAN TÜRKOĞLU

Lab

AI-assisted product execution lab.

I use AI to accelerate product execution — prototyping, documentation, testing, build. Product judgment remains human. The result is a smaller team moving with more structure, speed and production confidence.

JUDGMENT DECIDES WHAT SHIPS — THE LOOP CONTINUESFigmaTOOLProduct BriefHUMAN-LEDUX FlowHUMAN-LEDClaude CodeTOOLPrototypeOUTPUTQA / Test PlanOUTPUTIterationHUMAN-LED

AI accelerates execution. Product judgment decides what is worth building.

What AI helps with

  • Rapid prototyping
  • Documentation
  • Test scenarios
  • Product requirements
  • UI exploration
  • Technical briefs
  • Code-assisted iteration

What AI does not replace

  • Product judgment
  • Business context
  • Taste
  • Prioritization
  • User empathy
  • Stakeholder alignment

The tools draft. The operator decides.

Example workflows

input → output

Figma → Structured product site

Design files can flow through Figma MCP into coded, deployable builds; selected case-study assets and presentation workflows in this portfolio follow that same operating model.

Brief → PRD → flows → tasks

A vague client brief becomes a structured requirements document, user flows and a prioritized, sprint-ready task board — drafted with AI, decided by product judgment.

Design system → Coded components

Token collections map to theme providers and component code; the ePOINT v2 system is migrating into the live React Native app through exactly this path.

Product idea → Prototype → QA checklist

New flows ship with AI-drafted test plans and automated UI suites — features are regression-guarded before users ever meet them.

Selected experiments

real systems, not demos

Technical product architecture

Trading bot & control room

A TradingView-to-Bybit automation system: NestJS execution bot with a Next.js control-room panel for monitoring and intervention. Built AI-assisted, end to end — an exercise in productizing a technical system with real money-flow constraints.

QA at product scale

Automated UI regression

Automated and semi-automated UI regression suites for a live loyalty platform, paired with multi-persona review rounds. Findings land as structured cards on the QA board, not in chat scrollback.

Craft, scripted

Design-ops via Figma Plugin API

Token-driven variable-binding sweeps, sibling-aware WCAG contrast audits, layout linting and icon indexing across a large multi-hundred-screen system file. Quality at scale is tooling, not heroics.

Knowledge infrastructure

Product documentation systems

NotebookLM-backed project knowledge bases and structured work-logs that make product history queryable — requirements archaeology in seconds instead of afternoons.

The lab exists to make the product work faster and sharper — it is infrastructure, not identity.

Back to the product work