Skip to content
View jabluetooth's full-sized avatar
🏠
Working from home
🏠
Working from home

Block or report jabluetooth

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
jabluetooth/README.md

Hey there! I'm Fil Heinz 👋

Automation & AI Solutions Engineer · n8n / REST / Webhooks · TypeScript & Python

Portfolio LinkedIn Gmail GitHub


name: Fil Heinz O. Re La Torre
location: Silay City, Negros Occidental 🇵🇭 (Remote)
education: BS Computer Engineering @ TUP-Visayas
focus: Building automations & integrations that go live in days, not months

⚡ What I Do

I build practical automation and integration solutions - connecting CRMs, ops platforms, AI services, and databases into workflows that just run. I work in n8n, Python, and TypeScript, design REST/webhook/OAuth integrations, and put LLMs to work on real business processes (content generation, document tailoring, support triage, semantic dedup). I care about systems that don't block, don't lose state, and degrade gracefully when an upstream API misbehaves.

"Automate what a computer should do, surface what a human needs to see."


🔥 Featured Projects

🎯 Match - AI Job-Matching SaaS with an n8n Automation Backend   ·   live ↗

A thin Next.js app talking to a 6-webhook n8n backend over signed (x-webhook-secret) calls. Long-running work - scraping, LLM scoring, resume tailoring, PDF rendering - runs asynchronously in n8n while the app polls Postgres for completion, so the UI never blocks or loses state. Integrates JSearch, RSS feeds, NewsData.io, Groq, PDFShift, and Gmail; defensive LLM-JSON parsing degrades visibly instead of breaking.

n8n · Webhooks · Next.js · TypeScript · PostgreSQL / Prisma · Groq · Clerk · Vercel

📰 ZeroPress - Zero-Touch AI Content Pipeline   ·   live ↗

A 37-node n8n pipeline that runs every 6 hours (or on demand): scans 4 trend sources, dedupes and ranks, selects a topic with an LLM + brand-safety gate, researches it via parallel Google Search queries, writes a 1,500-2,500 word SEO article, generates a cover image, publishes via an authenticated REST API, emails subscribers, and logs every run. Fully autonomous, $0/month infra.

n8n · Groq · Hugging Face FLUX · SerpAPI · Supabase · Brevo · REST API · Next.js

🔍 Mimo - RAG-Powered Internal Knowledge Assistant   ·   live ↗

Retrieval → reranking → confidence-gated generation, with per-claim citations and an explicit refusal path instead of hallucinated answers. Measured against the live system, not a local mock: 100% retrieval accuracy (25/25) and 100% prompt-injection resistance (12/12) on its eval suite, with a documented before/after threshold tuning pass.

n8n · Groq · Qdrant · HuggingFace · JWT / RBAC · Postgres

🩺 Insight - AI Root-Cause Copilot for n8n Failures   ·   live ↗

Paste a failed n8n execution and get a plain-English diagnosis of which node broke and why, instead of reading raw execution JSON by hand. Can auto-install its own error-monitoring workflow onto a connected n8n instance via a narrowly-scoped, disclosed write-access model.

n8n · Groq · Next.js · PostgreSQL · Webhooks

💪 Se7en - AI-Coached Fitness App (RAG)

An offline-first React Native app whose AI coach uses retrieval-augmented generation: HuggingFace embeddings stored in Neon pgvector, retrieved per query and fed to Groq Llama 3.3 70B. Typed schema, custom chart library, state machines over booleans.

React Native · Expo · TypeScript · Firebase · Neon (pgvector) · Groq · RAG

🤖 Help Scout → AI Knowledge-Base Automation (client project, private repo)

A 35-node n8n workflow that turns resolved support tickets into reviewed KB articles: webhook intake → conversation extraction → LLM-drafted article + summary → human review in Airtable → auto-publish to Help Scout Docs API → Slack notify. Uses Gemini embeddings + Pinecone for semantic de-duplication so it never publishes redundant articles.

n8n · Webhooks · Groq · Gemini Embeddings · Pinecone · Airtable · Help Scout API · Slack


🛠️ Tech Stack

Automation n8n
Integrations REST Webhooks OAuth
Languages Python TypeScript JavaScript SQL C++
AI / LLM Groq HuggingFace Gemini Pinecone Qdrant
Data & Backend Postgres Supabase Prisma Drizzle Firebase Redis
Frameworks Next JS React React Native FastAPI TailwindCSS
Cloud & DevOps Vercel AWS Docker GitHub Actions

GitHub Stats Most Used Languages


🧭 How I Work

  • Ship fast, ship live - I scope, build, and deploy solutions that go into production in days, not quarters.
  • 🔌 Integration-first - REST, webhooks, OAuth, and signed server-to-server callbacks; I make systems talk to each other reliably.
  • 🛡️ Resilient by design - async orchestration, defensive parsing, auth & secrets handling, and graceful failure over silent breakage.
  • 📝 Documented & adoptable - I leave behind clear docs so the next person (or client) can run and extend the work.

Profile Views

Built with caffeine, curiosity, and a habit of automating the boring parts ☕⚙️

Pinned Loading

  1. bonny-ai bonny-ai Public

    AI-powered, database-driven portfolio. Chat naturally to explore my work, with responses that dynamically render UI components in real time.

    TypeScript 1

  2. se7en se7en Public

    AI-coached gym tracker with a rotating 7-day cycle. Offline-first React Native app with RAG-powered coaching insights (Groq + pgvector).

    TypeScript

  3. match match Public

    AI-powered career management dashboard that matches jobs to your resume, tailors it per posting, researches companies, preps interviews, and tracks every application via an n8n automation backend.

    TypeScript

  4. mimo mimo Public

    RAG-powered internal knowledge assistant. Retrieves the right passage and answers with a citation, or refuses rather than hallucinate. 100% retrieval accuracy and prompt-injection resistance on its…

    TypeScript

  5. zeropress zeropress Public

    Zero-touch AI publishing pipeline: a 37-node n8n workflow that researches trends, writes and illustrates SEO articles, and publishes them autonomously every 6 hours on $0/month infra.

    JavaScript

  6. what-to-do what-to-do Public

    AI app-idea-to-boilerplate generator. Describe an app in a sentence and get a full PRD, a recommended tech stack with reasoned alternatives, and a scaffolded project with a live in-browser preview …

    TypeScript