Open liveDecision-Support Application
LiveCan I afford to retire?
RetireCheck
See if your savings may last — with clear disclaimers.
Weidong Shi · AI Solutions Architect
AI in Action
Build → Validate → Improve → Document → Share. Personal production case studies in AI-assisted engineering — for business and technical audiences alike.
5 live applications · open any panel
Live apps
Five production apps — open the product, then read the story behind the build.
Open liveDecision-Support Application
LiveRetireCheck
See if your savings may last — with clear disclaimers.
Open liveLocal-First Wellness PWA
LiveSleepCheck
Wind down locally — soundscapes, stories, no account.
Open liveLocal-First Habit Coach
LiveHabitCheck
Honest weekly habits — no streak theater.
Open liveAdvisory Assessment Application
LiveReadiness Advisor
Eight dimensions, hard gates — advisory, not a stamp.
Open liveEnterprise Architecture Intelligence
LiveArchLens AI
Cited evidence from disk — refusals when claims don't resolve.
Method
Each app follows the same loop: ship something real, validate in production conditions, improve from evidence, document the decisions, then share the case study.
Proof
Architecture decisions, trade-offs, and delivery notes — still here, still linked from every live app.
Open liveLocal-First Habit Coach
LiveMissed days shouldn’t force fake checkmarks; people need honest recovery that keeps momentum real.
AI Applications · Recovery · Facts vs Coach · Privacy-gated AI
Open liveDecision-Support Application
LivePeople need a credible “will my money last?” answer without advisor-grade onboarding friction.
AI Applications · Domain architecture · Monte Carlo · Cloud delivery
Open liveLocal-First Wellness PWA
LiveWind-down tools often over-collect data; people need calm UX with a privacy-first architecture.
AI Applications · Local-first · Privacy · Mobile-first PWA
Open liveAdvisory Assessment Application
LiveShipping an AI demo is easy. Shipping AI that people can trust is much harder — teams need the right questions before deployment, not another checklist stamp.
AI Applications · Eight dimensions · Hard gates · OpenAI advisory narrative
Open liveEnterprise Architecture Intelligence
LiveArchitecture reviews fail when recommendations are unfalsifiable — hand-authored findings with no check that sources exist or say what the claim asserts.
AI Applications · Verifier gate · Evidence spans · Human approval
Capabilities
Senior architecture judgment applied to AI-enabled delivery.
Frame the problem, choose boundaries, and design systems that stay correct, private, and operable.
Ship cloud-hosted products with clear service edges, CI, and production judgment—not demos that stop at the UI.
Use AI assistants to accelerate delivery while keeping architecture, tests, and risk ownership with the engineer.
Decide what never to collect. Local-first and data-minimization when the product requires trust.
Featured insight
Judgment on the page — start here if you want the thinking behind the builds.
Unity Catalog did not block it. By default, catalogs can be accessible from every workspace attached to the same metastore, subject to privileges.
Your development workspace may be able to read production data. Unity Catalog did not block it. By default, catalogs can be accessible from every workspace attached to the same metastore, subject to privileges.
Watch & save
Short LinkedIn clips and frameworks — useful on their own, better when they lead you back to a case study.

Your development workspace may be able to read production data. Unity Catalog did not block it. By default, catalogs can be accessible from every workspace attached to the same metastore, subject to privileges.

AGENTS.md, rules, MCP, commands, skills, subagents, hooks, ignore — the seven files that decide whether Cursor knows your repo or guesses at it.

Confident hallucination, stale RAG, prompt injection, silent fallback, runaway agent — every one of them passes a demo. A demo proves the happy path, not that the system is trustworthy.
About
Based in the Greater Pittsburgh Region. Career spanning enterprise platforms, CMS/DXP, cloud delivery, and long-lived product systems — now documenting AI-enabled delivery through a personal lab.
Next step
Exploring an AI-focused role, or want a senior perspective on architecture and delivery? I’d enjoy hearing what you’re building.