How I work
Useful systems need human judgment.
I use AI-assisted tools to build and revise work, but the brief, the evidence, the review, and the decision to use an output still matter. These examples show both the resulting tools and their limits.
Working method
From context to a checked result.
I define what needs to happen, supply context, review a working version, test claims and behavior, and ask for changes where the result falls short. AI-assisted coding tools help implement the systems; my contribution is in directing and reviewing their use.
- 01Define the need
- 02Supply context
- 03Review the work
- 04Test the result
- 05Revise or use
Training planning and review
The need
Coaches and lifters need a usable connection between a plan, completed sets, progress, and video feedback.
My contribution
Peter supplied the training context and workflow requirements, reviewed successive versions, and checked whether the tools reflected the way sessions are planned and logged.
Method and result
An AI-assisted development process helped build and revise the connected Excel system and the Django application. Peter evaluated the interaction and requested corrections. The workbook links plans, session logs, and record boards. The capstone connects workout records with a video-review workbench.
Present limit
The application remains in development. Automated video tracking is experimental; it is not a validated motion-measurement service.
Content review and approval
The need
Professional posts need evidence, privacy review, and a human decision before publication.
My contribution
Peter defined the approval boundaries and the requirement to trace claims back to authorized material.
Method and result
A local Python and SQLite workflow organizes sources, drafts, review gates, and an audit history. AI assistance was used during development and testing. The tool catches unsupported or private content before a final approval decision and records state changes.
Present limit
Its publishing adapter is simulated. It does not run live LinkedIn posting or a client campaign.
Website delivery and maintenance
The need
A useful site must turn a brief into clear pages, preserve existing work, and remain reliable across devices.
My contribution
Peter set business requirements, reviewed visual and editorial decisions, supplied corrections, and determined what belongs in the public record.
Method and result
AI-assisted implementation and review supported copy, code, responsive checks, source preservation, and release preparation. This portfolio joins project cases, services, and the original FAST Strength inquiry site on one domain.
Present limit
The original strength storefront remains inquiry-based. Website and code checks are evidence of a reviewed process, not a claim that Peter independently wrote every line.