I do not sell AI as magic. I use it as leverage. My work is public, reviewable, and preserved.
| Area | Proof |
|---|---|
| Open-source production work | Security-sensitive PRs in openclaw/openclaw-windows-node |
| Evidence preservation | Built OTSkit — OpenTimestamps toolkit with MCP integration for timestamping and preserving digital artifacts on Bitcoin |
| Local-first privacy | ContextSafe — local anonymization for legal/personal documents |
| Practical public contributions | DeclaRenta tax automation |
| Project | What it proves | Stack | Links |
|---|---|---|---|
| OpenClaw | C#/.NET desktop security under public review | C#, .NET, WinUI 3, WebView2 | upstream / fork |
| OTSkit | OpenTimestamps toolkit + MCP integration + evidence preservation | TypeScript, OpenTimestamps, MCP, SHA-256 | org / core / MCP |
| ContextSafe | Local-first privacy tooling with real legal-tech use case | Python, FastAPI, React, spaCy, Presidio | repo / site |
| DeclaRenta | Focused TS/React contributions with fiscal impact | TypeScript, React | repo |
Desktop & Systems
Web, API & Data
Security, Preservation & Automation
My role is to:
- Define the problem and threat model
- Split large work into reviewable slices
- Check generated code against edge cases
- Reject output that does not survive scrutiny
- Respond to review with concrete fixes
- Leave behind artifacts that can be inspected later
Focused on verifiable work. No inflated project list, no dead SaaS claims, no unverifiable noise.