Continuity, thoughtful engineering, and evidence people can check.
I'm S.J., the independent creator behind Estra Logics. I follow patterns, ask practical questions, and turn them into tools and experiments. My work brings together software investigation, continuity research, and creative systems.
Explore the portfolio · Read the work record · Estra Logics home
Estra Logics is the platform. CML — Continuity Markup Language — is its underlying language and engine. I develop CML as an enhancement to existing software: a way to describe what must remain true, check proposed changes, and preserve a record of the result. Python, JavaScript, and other host tools carry out their own work alongside it.
Fermata is the working method around that engineering: observe, pause when a condition is unresolved, check the evidence, and continue when the conditions are met.
The question running through it all is simple: what must remain true when work changes hands, tools, or form?
| Project | What it explores | Stage |
|---|---|---|
| CML Local Alpha | Static continuity checks, accepted-state recovery, and offline verification receipts | Alpha |
| CML Trust | Traceable observations, evidence hashes, and preserved revocation history | Research prototype |
| Transparent Glasshouse | A shared interface where people and agents can see state and actions | Public demonstration |
| Pattern Atlas | Timelines and relationship maps with visible source notes | Browser prototype |
| Engineering Work Record | Dated investigations, patches, validation results, and handoffs | Public record |
Each project's documentation describes what it implements and what its checks cover.
Microsoft Semantic Kernel #14454 — preserving Pydantic field descriptions in Python schemas. Merged October 27, 2026.
The upstream pull request preserves the change, review, and contribution history.
My research interests include mathematical patterns, harmonics, music and visual continuity, and the structure of historical NASA programs. I explore these through source reading and bounded experiments, keeping observations and interpretations distinct.
I work with AI as a technical partner. I guide the questions and direction; the work stays reviewable through source, tests, and preserved evidence. Corrections remain part of the record.
Open to thoughtful collaboration and contract work in software investigation, validation tooling, engineering coordination, and technical documentation.
Human-led · AI-assisted · Evidence-bound