Adaptive productivity, database auditing, workflow automation, and the infrastructure that makes AI-assisted work traceable rather than magical.

Open-source, privacy-conscious processing for personal ChatGPT export archives: a staged local pipeline (intake → Phase 1 reports → verification → Phase 2 selective staging → structural indexing) that runs entirely on your own machine, with redacted outputs, provenance-hash checks and fail-closed verification. Released as CI-tested v0.1.0 under Apache-2.0 — tests and the demo run on synthetic fixtures only, never real data.

🧰 Open source: adrianadroid-hash/archive-intelligence

Adaptive productivity built around variable human capacity — modeling energy, executive-function friction, and behavioral state instead of pretending every day is the same.

📄 Case study (documentation only — source remains platform-hosted): adrianadroid-hash/mindflow-case-study

A structured database-audit system turning messy operational data into scored, risk-classified findings with auto-fixability assessment and remediation paths.

📄 Case study (documentation only — source remains platform-hosted): adrianadroid-hash/nexustask-case-study

Runnable automation experiments showing the progression from manual operational work to deterministic workflows, APIs, and AI-assisted systems — built on synthetic data, with failures and repairs recorded, not just polished outputs.

🧪 Open lab: adrianadroid-hash/ai-automation-lab

Product thinking · Systems design · Workflow automation · Data quality · AI-assisted development · Operational problem solving

• Personal portfolio — a Gravitational Information System–style interface for presenting evidence-backed work (in development)
• AI Automation Lab — next exercises: automated tests, synthetic catalogue-audit pipeline, documented failure/recovery paths

Case studies are documentation-only and clearly labeled as such. Where source is not available, the repository says so — no reconstructed or substitute code.