Pulling a Long-Lived Key Out of the Client: Governing an Inherited Upload Chain
How I took over a mini program with a long-lived signing key shipped in the frontend, restored uploads first, then moved signing to the backend in a second phase.
Case studies
Projects
Problem, constraints, trade-offs, and verified results.
Trustworthy Data First: Permission and Interpretation Boundaries on a Behavior-Data Chain
Why AI should get a controlled, auditable data outlet instead of a database account, and where its interpretive authority has to stop.
Two Systems, One Person: Why I Did Not Merge the User Tables
Why weak coupling, explicit failure postures, transactional outbox, and nightly reconciliation were the cheaper and more honest design for linking one person across two independent systems.
From a List to a Map: Turning a Bid Archive into a Verifiable Industry Asset
How a hand-compiled five-year award archive became a governed industry map — and why the value of a data asset lies in verifiability, traceability, and governability rather than volume.
After AI Says “Done”: Contracts and Independent Verification for Real-World Delivery
How explicit task contracts, independent review, and fresh-eye reproduction turn “the AI says it’s done” into portable, checkable evidence.
Where AI Belongs in a Business Data Pipeline
A tender-intelligence pipeline shows where models help, where deterministic code and named people must stay accountable, and why a finished pipeline is not validated demand.