Can an off-the-shelf tool do the job?
If the integrations already exist, a workflow tool is faster and cheaper to run. If the job needs custom logic, heavy data work or answers from your own documents, I write the backend.
I automate the busywork, and write
code when a tool can't.
Full-stack and AI automation engineer in Albay, Philippines, helping U.S. teams answer leads faster and stop moving data by hand.
Sample walkthrough: a new email is classified, the knowledge base is searched, a reply is drafted for a person to approve, and the CRM record is updated.
- ~4 hours a day of inbox triage removed by a self-hosted AI operations platform.
- 200+ institutions use CAFE, the compliance portal I built for CHED Region V.
- 100+ MBA users supported on a U.S. client's AI coaching pilot.
Automation where it fits. Custom code where it doesn't.
Every job starts with the same question, and the answer decides how I build it.
Yes: automate it with tools
Fast to ship and cheap to run when the integrations already exist.
- Inbox and inquiry triage, with a person approving replies
- Voice agents that answer inbound calls around the clock
- Lead routing and cleanup in HubSpot or Monday.com
- n8n
- HubSpot
- Power Automate
- Vapi
No: build it custom
For business logic, volume or private data that a workflow canvas can't hold.
- Web platforms and portals in Laravel and Vue
- Python APIs and AI pipelines that answer from your own documents
- PostgreSQL data models that keep each client's data separate
- Laravel
- Vue
- FastAPI
- PostgreSQL
Either way, I deploy, host and maintain it
Hardened Linux or Azure, Docker, SSL, backups and DNS, so it stays online after launch.
Not sure which path fits? Send me the process and I'll tell you
Shipped and running in production.
Open them, click through them, or read the code.
AI Customer Operations Platform
Full-stack and AI engineer, 2026
Captures inbound leads, sorts incoming email, and drafts replies from the company's own knowledge base. A person approves every reply before it sends.
Removed about 4 hours a day of manual inbox work, with no per-message AI cost.
Stack: Laravel 13, Vue 3, n8n, Qdrant hybrid search, Docker Compose
CAFE compliance portal, CHED Region V
Sole developer, OJT intern, 2026
Where higher education institutions across the Bicol region submit compliance forms, certificates and reports.
Live for 200+ institutions, with files on AWS S3 so the server never fills up.
Stack: Laravel, Filament, AWS S3, Sanctum SSO, MySQL
Workforce operations platform
Full-stack developer, contract, 2026
Shift scheduling and compliance for several companies in one system, with conflict-free booking.
Calendar changes reach every user instantly, and each company's data stays separate.
Stack: Laravel 13, Vue 3, Redis Horizon, WebSockets, PostgreSQL
Application tracking and monitoring system, PCA Region V
Full-stack developer, 2025
Replaced paper tracking with an online system, SMS alerts and a map for verifying project sites.
Runs on a hardened Linux server with automatic SSL and backups.
Stack: Laravel, Leaflet.js, Nginx, DigitalOcean
Also on GitHub: a Bash toolkit for BIND9 DNS that cut zone setup from 30 minutes to under 2 per domain. See all repositories
Client work under NDA
Two U.S. engagements under NDA, so no names or screenshots. Here's what I built for each.
Legal marketing and referral network
Senior AI Automation Engineer, Aug 2026 to present
- Referral tracking app
- An internal app that routes referrals to partner firms and tracks each response, with automated SMS reminders and CRM sync.
- Sales follow-up from call transcripts
- An AI workflow reads discovery-call transcripts, pulls out objections and buying signals, and drafts a tailored follow-up. Prospects who already booked are skipped automatically.
- HubSpot pipeline and lead routing
- Rebuilt lifecycle stages from lead to closed-won and added routing rules so every inquiry lands in the right pipeline before it reaches the CRM.
- Production AI backend on Azure
- Moved a workflow-canvas prototype to a FastAPI backend on Azure, keeping n8n only where it fits.
- B2B lead enrichment
- A pipeline that researches and scores prospective accounts, delivering deduplicated, CRM-ready imports.
Executive coaching and leadership firm
AI Applications Support Specialist, 2026
- AI coaching platform and MBA pilot
- Supported the platform's AI assistants and user access through a university pilot with 100+ MBA users.
- Microsoft 365 operations automation
- Power Automate and Monday.com flows plus a custom operations agent move documents, forms and schedules without manual handoffs.
- AI-powered Outlook inbox screening agent
- Built and deployed a Microsoft Graph integration on Railway that monitors Outlook, classifies likely solicitation emails with AI, and routes high-confidence matches to a review folder, keeping trusted and ordinary messages in the inbox.
Need something similar? Email me. I'm happy to sign your NDA before we talk details.
Experience, education and tools.
Experience
Recent
- Aug 2026 to now
Senior AI Automation Engineer
Legal marketing and referral network, U.S. (remote)
- 2026
AI Applications Support Specialist
Executive coaching and leadership firm, U.S. (remote)
- Apr to Jun 2026
Full-stack developer, contract
Workforce operations platform
- Feb to May 2026
OJT intern, HEMIS Unit
CHED Region V, Legazpi City
Earlier
- Apr to Nov 2025
Full-stack developer, capstone
PCA Region V application tracking and monitoring system
- May 2025
Election technical support staff
COMELEC Malinao, Albay
- May 2022
Network infrastructure intern
DCTV Cable Network Broadband Services, Tabaco City
Education
- 2026
BS Information Technology, Magna Cum Laude
Bicol University College of Science
- 2021
Computer Systems Servicing NC II
TESDA national certificate
Toolkit
- Automation and AI
- n8n, HubSpot, RingCentral, Power Automate, Microsoft Graph, Monday.com, Vapi, Qdrant, Azure AI Foundry, Ollama
- Applications
- Laravel 13, Vue 3 and Inertia, FastAPI, SQLAlchemy, PostgreSQL, MySQL, Redis, Tailwind CSS
- Infrastructure
- Azure, AWS S3, DigitalOcean, Railway, Tailscale, Docker Compose, Nginx, Linux, BIND9 DNS, GitHub Actions
Questions clients ask first
- Are you available for part-time, contract, or retainer work?
- Yes. I take project-based contracts, weekly sprints, and monthly retainers. I work remotely from Albay, Philippines, with daily overlap for U.S. Eastern, Central, and Pacific hours, over Slack, email, or scheduled calls.
- How do you decide between an automation tool and custom code?
- If existing integrations and triggers solve the problem, a tool like n8n or HubSpot is faster and cheaper to run. When the job needs custom business logic, heavy data work, high volume, or AI answers grounded in your own documents, I build a Laravel or FastAPI backend instead. On a U.S. engagement I moved an AI prototype off n8n onto a FastAPI backend on Azure. I pick whichever will be easier to maintain.
- Can we sign an NDA before sharing details?
- Yes. All my client work runs under mutual confidentiality. Your processes, data, client records, and code stay private.
- What have you shipped for clients under NDA?
- For a U.S. legal marketing and referral network I built an internal referral tracking app with SMS reminders and CRM sync, AI sales follow-up from call transcripts, and a production FastAPI backend on Azure. I also supported an AI coaching platform used in a university pilot with 100+ MBA users. Client names and internal details stay private.
- Can you set up voice agents, lead qualification, and CRM routing?
- Yes. I set up Vapi voice agents that answer inbound calls, qualify the caller, book a calendar slot, and save the call record and transcript to HubSpot or Monday.com.
- Can you build AI automation and RAG systems?
- Yes. I shipped a self-hosted AI customer operations platform with 8 production n8n workflows and a RAG pipeline using hybrid search (dense and BM25), Reciprocal Rank Fusion, and cross-encoder reranking on Qdrant, validated at 100% Recall@5. A person approves every reply before it sends, and the system removed about 4 hours a day of manual inbox work.
- What happens after launch?
- I stay on for maintenance: updates, database backups, server hardening, and domain and DNS management on GoDaddy, Namecheap, or Cloudflare.
- How quickly can we start?
- Usually within 48 to 72 hours. We start with a short call to map the bottleneck and your current tools, agree on deliverables and milestones, then I start building.
- Do you also take operations or virtual assistant work?
- Yes. Alongside development I take operations and automation support: process automation, data cleanup, reporting, and internal dashboards. I can remove repetitive manual work, and I am open to ops or VA roles.