Builds · For hiring managers + clients · 2026

Show the work before the results land.

Senior Operations Builder + Leader. 20+ years across five industries. Bilingual. This page is built first for hiring managers evaluating a senior ops candidate (and second for clients evaluating Analytic Gator). Each entry is a working system: what was built, how it was architected, what's been learned. Five live, two in build for first paid clients. The trust signal isn't outcomes yet; it's the work.

If you're a hiring manager: scan the build statuses for live systems and read Career Desk first. If you're a client prospect: see analyticgator.ai for engagement tiers.

Live · v3.1 · Audit-and-fix loop in production

Brand Desk

Managed agent · Surface Briefs + audit-surface · Semantic checks

The system that maintains the brand. Surface Briefs catalog covering 22 public surfaces + the /audit-surface agent + AUDIT-RULES + maintenance protocol. Catches drift before it cascades, auto-fixes mechanical patterns with safety constraints, reasons about register inconsistency that regex can't see. Context architecture applied to my own brand — eating the dog food.

Claude Code skill Markdown catalog Auto-fix + auto-commit Semantic checks (S01) HITL voice gate
Read the build →
Pre-launch · v0.5 · First applications shipping soon

Career Desk

Job-search agent · LLM Wiki Paradigm · HITL

My own job-search system, running every morning. Discovery agent finds roles on residential IP, fit-scoring qualifies them, the LLM Wiki Paradigm holds compounding company research, and Alfonso approves every output before anything moves forward. Three-deliverable application packages. Senior architectural choices documented in 10 ADRs.

Supabase Claude API Puppeteer agent LLM Wiki HITL
Read the build →
Pre-launch · v1.0 · Awaiting A2P approval

Family Care Desk

SMS · Voice · Web · Bilingual · Private

A private bilingual care coordination system for one Mexican-American family, shipped in 11 days. Text the line with anything you'd tell Mom: appointments, medications, photos. The system files them in the right place and surfaces them when the family needs them. End-to-end: 10 tables, 5 Edge Functions, 2 languages, 3 input modes.

Supabase Twilio A2P Claude API Bilingual systems Family ops
Read the build →
Pre-launch · Phase 1 complete · First piece shipping soon

Content Desk

Marketing OS · Claude Code skills · Voice gate

A marketing operating system implemented as Claude Code slash commands instead of an app. Five skills (story → write → voice-check → publish → archive) with HIL gates at every step. Every session produces two outputs: the content and the system training. Voice gate enforces a 15-point quality bar before anything ships.

Claude Code Open Brain Voice integrity Build in public
Read the build →
Site live · Pre-launch · First dispatch shipping soon

HumanOperators.ai

Personal news feed showcase · Newspaper structure

A personal news feed for the AI era, structured like a newspaper. Not a category publication, not an audience-building flywheel. The site IS the showcase: what one operator's branded, public, contextual feed looks like. Powered by the Content Desk pipeline; fed by Capture.

Editorial design Vercel Content Desk pipeline ⟨Operator⟩
Read the build →
Live · v1.0 · Deployed Apr 24 · Captures flowing

Capture

One-day build · Three files · Input layer for everything

The smallest build in the portfolio, the deepest thesis: capture friction is the actual bottleneck of content production. A bookmarklet + an iOS Shortcut + a Supabase captures table. Three files. Deployed in a day. Now feeding Open Brain, Content Desk, Career Desk bullets, and HumanOperators dispatches.

Bookmarklet iOS Shortcut Supabase Input architecture Personal tool
Read the build →
In build · First paid AG client · May to August 2026

Pinks Desk

Tier 2 Operator Desk · Pink's franchise · Bilingual

First paid Tier 2 client engagement under Analytic Gator. A franchisee operating brain for Pink's Window Services, sitting on top of the ResiBrands corporate platform. Operator's daily brief, bilingual crew dispatch (Spanish-default), customer-facing recurring activation, leadership dashboard. The lead Tier 2 case study under construction.

Supabase Twilio Bilingual operations Franchise ops AG Tier 2
See the architecture sketch →
In build · Tier 1 summer pilot · June to August 2026

Teacher Desk

Tier 1 Solo Desk · Public school teacher · Summer pilot

First Tier 1 client engagement under Analytic Gator. A single-surface Desk for a public school teacher: curriculum context, student context, parent communication, principal asks, last year's lessons. Built over the summer, ready for the fall semester. The lead Tier 1 case study under construction.

Education Bilingual Solo professional AG Tier 1
See the AG strategy →

Why pre-launch? Established firms post outcome stories. Builders post build stories. The unconventional move is publishing before the metrics exist. That's the trust signal, not a weakness. Each page below has a "What success looks like" section instead of a "Here are the numbers" section. Once the runs complete, those pages will get updated with what actually happened.

Why the asymmetry? Six multi-week systems, one one-day tool, and one one-day-shipped agent that lives across all the others. They look incommensurate; they aren't. They're operating at different scales for different reasons: Career Desk needs a real codebase because it handles hundreds of jobs; Capture needs three files because the entire thesis is capture cost approaches zero. A capture system that took weeks to build would refute its own argument by existing. Brand Desk shipped in one stretch because the system is mostly markdown + a Claude Code skill — context architecture doesn't need infrastructure to be real. Builders ship at multiple tempos. The portfolio is stronger when it shows them all.