About
I analyze messy workflows, find the real constraint, and turn it into a working system.
I'm Luke, an AI Product Builder and Technical Product Manager based in Ontario. I build operational software end-to-end in unfamiliar domains: learn the workflow fast, reduce risk before code, and ship the smallest reliable system real users adopt.
Strong across product scoping, system design, Supabase/RLS architecture, AI-assisted execution, and production QA.
Why This Portfolio Exists
I use Craftmindship to show how I turn messy real-world workflows into shipped software.
The projects here are not mock case studies. Cadence came from working with a real music teacher trying to coordinate lessons, parents, practice notes, and scheduling. Puppy Program OS came from observing the operational load behind guide dog fostering. Multi-Agent Workflow documents how I use AI tools as a disciplined build system, not random prompting.
The common thread: find the constraint, design the system, ship the workflow, then prove it with real usage. Every project starts with planning documents, not code. AI tools execute against specs, not vibes.
Professional Experience
Product Builder & Technical Product Manager
Cognito Technology Canada Inc. · Aug 2023 – Present
Translated ambiguous founder briefs into production-ready architecture across a portfolio of 20–30 ventures in SaaS, fintech, logistics, and edtech. Owned product scoping, schema design, implementation planning, QA review, and deployment handoff across multiple concurrent MVPs.
Organization Design
GEEK Up · Aug 2020 – Sep 2021, Jul 2022 – Dec 2022
Designed organizational systems and operational workflows for a growing technology company. Redesigned competency frameworks, contribution models, and performance review processes used across engineering and design teams.
Learning Design Expert
Topica Edtech Group · Nov 2021 – Jul 2022
Analyzed platform workflows and operational data within a large-scale edtech environment serving 100K+ learners. Mapped LMS/CMS data flows and identified integration gaps that informed product improvement roadmaps.
Education & Certifications
B.Eng. Engineering Physics — Ho Chi Minh City University of Technology (Bach Khoa)
Google Data Analytics · IBM Data Analyst · Google Project Management
Technologies
Web: React · TypeScript · JavaScript · HTML/CSS · Tailwind
Data: Python · SQL · Power BI
Backend & Infra: Supabase · PostgreSQL · RLS · Edge Functions · Vercel · GitHub
AI workflow: Claude · Claude Code · ChatGPT · Codex
QA: Playwright · Manual mobile QA
Project & Workflow: Jira · Notion · Obsidian · Draw.io · Whimsical
Looking For
How I Build
Product Discovery
System design and planning in Claude Chat — 80% planning, 20% execution. Planning artifacts (MD files) drive every build session as agent context.
Database Architecture
Supabase + PostgreSQL + row-level security policies. Access control enforced at the database layer, not the application layer.
Frontend Execution
Claude Code for precision implementation. Codex for targeted patches and refactors.
QA
Playwright browser testing for critical flows. Manual mobile QA for responsive and edge-case coverage.
Deploy
Vercel, auto-deploy from main branch. Ship continuously, validate in production.
Some production and company work cannot be public because it contains client, company, or real-user context. This portfolio focuses on public-safe case studies, shipped artifacts, and product reasoning rather than exposing private repositories.