I run courtroom operations at Santa Barbara Superior Court: courtroom clerks, court reporters, and court interpreters. Ten years inside California courts taught me how systems serve people, and exactly where they fall short.
In July 2025 I started building AI tools from the inside of the institution, not at it. I direct, AI generates, I validate. Nothing ships until it holds up. I also speak to court leaders about AI adoption (NACM 2026, Los Angeles Superior Court).
Off duty: Supernote devotee, Lakers lifer, and certain that earl grey hot is the correct order.
🛰️ Open to: roles where civic experience meets AI building. Prototyping, AI enablement, public-sector AI, or a job that didn't exist a year ago. AI-assisted building gave me wings. I want to use them where institutions and the people they serve need a translator.
🎓 Credentials: AWS Certified AI Practitioner · AWS AI/ML Scholar, Agent Engineer Nanodegree (2026) · Certified Court Manager (CCM) · MPA, Claremont Lincoln University · Paralegal Certificate, UCSB · BA Liberal Studies, CSU Long Beach
🟢 Live · Methods · Demo video · Repo
🟢 Live · Demo video · Repo: perseus-clew
🟡 Repo · PRD · Office Keys, its predecessor
🟡 Follow the build on dev.to · Running on Shako · Repo coming soon
Full fleet roster (click to expand)
| Vessel | Mission | Status |
|---|---|---|
| Lumen Clew | Code health scanner. Awareness, not shame | 🟢 Live |
| Ariadne Clew | Turns messy LLM chat sessions into preserved reasoning | 🔵 |
| Heimdall | Tests whether an AI gateway's access controls actually hold. Evidence for a sponsored dev.to piece | 🔵 |
| Daybreak | Always-on AWS agent that tracked everything I ship. Mk I of Observator | 🔵 |
| Hermes Clew | Agent readiness scanner, the proof of concept that became Agentis Lux | 🔵 |
| Anton Clew | Policy CLI that classifies AI file changes as DENY, NEEDS_APPROVAL, or ALLOW | 🔵 |
| README Clew | The first version of README Clew 2. I run it on mine. It found things | 🔵 |
| TicketGlass | Transparency for IT support tickets | 🔵 |
| Athena Clew | Autonomous debugging agent on Gemini 3 | 🔵 |
| Argus Clew | Agent-readiness review on every AWS commit | 🔵 |
| Clew Quest | Open source project discovery matched to your learning goals | 🔵 |
| Vigil Crest | Telegram correspondent that watches the DEV challenge feed | 🔵 |
| Clew Forge | Lightweight framework for shipping fast without losing clarity | 🔵 |
| Propina | USD/MXN tip calculator. One prompt, eight minutes, live | 🟢 Live |
| 4 GB and a Dream | How far can a 4 GB Chromebook go? Four builds, nothing installed, every commit pushed from the chat | 🟢 Live |
| Necesito 🍵 | My wife was sick on vacation, so I built her a bilingual papel picado page. One tap texts me what she needs. Its story made DEV's Top 7 | 🟢 Live |
| Steep 🫖 | Tea leaf reading for your repo. DEV April Fools 2026, Best Use of Google AI | 🔵 |
| Office Keys | Gamified typing practice for workforce digital literacy | 🔵 |
| Sprint-Kit | PM skills for middle schoolers. Zero data collection | 🔵 |
| Proof of Scoville 🌶️ | On-chain salsa tracker. Yes, really. No regrets | 🔵 |
| Bench Power: Lakers 🏀 | Data storytelling on bench impact | 🔵 |
🟢 Live · 🔵 Docked (repo, run it yourself) · 🟡 In drydock
More away missions (click to expand)
| Mission | What it does |
|---|---|
| Permission to Approach the Firewall | Cybersecurity awareness program for court staff |
| 12 Angry Rows | 11 years of jury utilization data, benchmarked against NCSC standards |
| Write of Access | AWS Textract on public court forms, with honest notes on what it can't do |
| Hung Jury | The blocker was data quality, not Python. Sometimes the finding is the deliverable |
| Panoptes Scout | First-pass data quality briefing agent for tabular datasets |
💎 7 Community Gems. Gems are DEV's human-curated picks. Trusted curators hand out a limited number each day to spotlight work they think deserves a wider audience. No algorithm, just people with taste.
| 💎 Picked as a gem | Date |
|---|---|
| Ghost Rovers: I Called 12 Machines NASA Left Behind. One Picked Up. | Nov 2026 |
| AI Took Me to Three Stages. I Still Don't Like Public Speaking. | Oct 2026 |
| My Agents Never Get Tired. I Do: On Satisficing | Sep 2026 |
| Block Zero: Oh, no! Claude, Kiro and I over-engineered the throwaway. | Sep 2026 |
| 💬 Plus 3 gems for comments, because the conversation counts too |







WeCoded Winner · Best Use of Google AI · Top 7 · JavaScript Author of the Week · 💎 Community Gems · 24-Week Streak
I don't write code from scratch. I give creative direction and validation: AI tools generate, I compare outputs, pick what works, and stitch it into something that holds up.
- Occam's Razor. Solve one problem completely before adding the next.
- Single Responsibility. Every component does one thing well.
- Deterministic structure, AI flavor. Code decides. The model describes.
- Name the limits. What a tool can't do goes in the docs, not under the rug.
- Operational Thinking. Ten years in courts means I know implementations have consequences.
This page redraws itself daily from the DEV and GitHub APIs. Last sensor sweep: 2026-11-20.


