Product engineer · antifragile systems

Ryan Gerard Wilson

I reverse-engineer brittle systems and rebuild them for production.

I read legacy Python, SQL, logs, operator corrections, and dirty data until the real rules are explicit. Then I replace the mess with smaller, inspectable software: decision engines, internal platforms, data pipelines, and automation another engineer can change without guesswork.

Shipped: a zero-downtime decision-engine cutover, a configurable internal platform, three full-stack Laravel products, and Go/Python tools built for the terminal.

Engineering method

Read the real system. Recover the rules. Replace the mess without breaking production.

Trace

Read what production actually does.

Legacy code, schemas, queries, logs, operator corrections, and edge cases outrank the old diagram.

Specify

Turn recovered behavior into constraints.

Make the data contracts, business rules, invariants, failure states, decision boundaries, and rollback paths explicit.

Simplify

Collapse the machinery.

Delete dead paths and repeated code. Pull variation into configuration only when the underlying shape is genuinely shared.

Cut over

Prove it under real load.

Validate against production evidence, ship incrementally, instrument failure, preserve operator recovery, and leave code another engineer can change.

Constraints

The work is not done if only its author can operate it.

No black-box model over a source of truth nobody can name.
No rewrite before legacy behavior and production edge cases are recovered.
No abstraction that merely hides duplicated business rules.
No cutover without validation, rollback, and operator recovery.
No system whose safe operation depends on its original author.

Contact