B2B SaaS built for a GTM client at Xponential7. A demand-gen report generator powered by a multi-stage LLM pipeline — ingests campaign data, runs structured AI analysis across multiple stages, and outputs actionable demand gen reports. FastAPI + Next.js + GCP. Live in production, client uses it daily.
Client projectA multi-agent workflow engine for turning complex tasks into structured, adversarially tested, decision-ready outputs. 7-layer architecture: Intake → Workflow Selector → Expert Graph Builder → Execution Engine → Decision Council → Ledger/Memory → Meta-Review. Includes full Skill Spec, Engineering Spec, and JSON schemas for all artifact types. Designed for Claude Code dynamic workflows.
LLM-powered business operating system — not a wrapper. A unified tool to fully operate your business: research, content creation, website and marketing material design, and anything else. Has all your business context, market intel, competitor data, and client knowledge built in. Work in progress.
Proprietary account intelligence layer. Monitors job changes, buying role movements, and company structure signals. Routes intent data to CRM and outbound triggers. The ABM data backbone — account scoring, buying role mapping, engagement monitoring — structured to plug into any sales workflow.
Custom research agent built in Claude Code. Automates competitive and prospect intelligence workflows end-to-end. Pulls structured account intelligence — stack, signals, relevant context for outreach. Compressed 30-minute manual research to under 60 seconds with structured, usable output.
TAM mapping and lead qualification system powered by Clay, connected to DataforSEO, custom scrapers, and dozens of additional APIs. SEO growth signals are one of many buying intent proxies — enriched with job-change data, tech stack signals, funding events, and any other signal a great GTM engineer pulls in. Structured for reuse across markets and campaigns with clean export to sequencing tools.
Multi-step automation pipelines built for client delivery. Cross-tool synchronization, webhook handling, CRM data routing, enrichment pipelines, trigger-based sequence launch. Every workflow documented and handed off — built to run without me in the loop.
Planning to launch a community for business and agency owners who use Claude Code, Codex, and other AI tools daily. Focused on best practices, benchmarks, hands-on testing, and sharing workflows, setups, and toolchains. For practitioners who live in these tools, not spectators.
An LLM-powered business operating system — not a wrapper. A unified environment where you fully operate your business: from research to content creation, website and marketing material design, and anything else. It holds all your business context, plus that of your market, competitors, and clients. Built to be the single place you run your company from.
Full-stack B2B SaaS built for a GTM agency client. A multi-stage LLM pipeline that ingests campaign data, runs structured AI analysis, and outputs actionable demand generation reports. FastAPI + Next.js + GCP. Live in production, used daily by the client. Not my own IP — built and delivered as a contractor engagement.
A comprehensive research and audit automation system that does market research, competitor analysis, customer research, and deep auditing of websites, offers, and marketing strategy — including sales funnels, customer journeys, and everything that can be found on a website or extracted from uploaded business and GTM documents. Works for your own business or for clients, and goes as deep as the user wants.
Powered by dozens of external APIs, multiple scrapers (Firecrawl, Tavily, and more), database connectors, and structured extraction pipelines. The end output includes audits for everything requested, plus new assets: better offer suggestions, Voice-of-Customer research, rewritten website copy, competitive positioning analysis, and much more.
A structured workflow engine that treats every complex task as a temporary expert organization: assembled for a single mission, run with discipline, and leaving a full audit trail. Built for Claude Code's dynamic workflow model — parallel subagents, adversarial red-teams, multi-round critique loops, and a senior synthesis layer that resolves tradeoffs before anything ships.
The framework ships with a complete Skill Spec, Engineering Spec v1, JSON schemas for all 11 artifact types (Mission Brief, Expert Map, Workflow Plan, Evidence Index, Decision Ledger, Red-Team Report, and more), and prompt templates for every phase. Designed to be installed directly into a Claude Code repo.
I build the thing, not a recommendation for the thing. Every engagement ends with working systems, not decks.
Claude Code in full auto-approve mode is my primary build tool. I ship in days what most quote in weeks. Documentation quality doesn't drop — the timeline does.
Everything I've built, I've run in production. I know the failure modes — deliverability gaps, CRM data model issues, signal routing edge cases — because I've fixed them.
Every system comes with documentation your team can run without me. The goal is always for me to become unnecessary.
Based in Cebu. Work EU business hours. Structured daily handoffs mean your team wakes up to progress, not status requests.
I work best with a defined outcome and timeframe. 3–6 months, clear deliverables, measurable output. That's the environment where I consistently over-deliver.