Building scalable systems from database to deployment Seeking internships, meaningful projects & ambitious teams
I architect systems that scale. Whether it's designing resilient APIs, optimizing database queries, or orchestrating microservices—I obsess over performance, reliability, and clean code.
My philosophy: Production-first. Ship fast. Learn faster.
const Developer = {
role: "Full-Stack Engineer (Backend-Heavy)",
expertise: {
backend: [
"REST APIs",
"System Design",
"Database Architecture",
"Real-time Systems",
],
ai: [
"LLM Integration",
"RAG Pipelines",
"Code Analysis Tools",
"Provider-Agnostic Solutions",
],
cloud: [
"Docker",
"Kubernetes",
"Cloudflare",
"Vercel",
"Cloud Optimization",
],
},
languages: [
"TypeScript",
"Python",
"Rust",
"Go",
"C/C++",
"Java",
"C#",
"JavaScript",
],
seeking: ["Internships", "Open-source", "Co-founder opportunities"],
response: "24 hours guaranteed",
};A two-player competitive time-tracking web app with live session tracking, gamified scoring, and side-by-side comparisons, built for mutual accountability and friendly competition between exactly two people.
- ✨ Live session tracking — Clock in/out with pause/resume, stop and save, and a midnight-to-midnight day timeline
- ✨ Gamified scoring — Crown scores, streaks, and archetype badges (Early Bird, Night Owl, Birb)
- ✨ Side-by-side comparison — Seven stat rows, dual day timelines, and 35-day activity grids
- ✨ PWA-ready — Installable on mobile and desktop with a warm parchment-themed UI
Stack: Next.js 16 · Supabase · TypeScript · Tailwind CSS · Drizzle ORM · Zod Website: time-tracker-ivory-two.vercel.app 🔒 This is a private repository. Please contact me for more details about this project.
A full-stack direct-booking platform for a fictional eco-resort in El Nido, Palawan, featuring an AI-powered concierge named Maya with RAG retrieval and a complete owner CRM.
- ✨ Zero-commission booking — 4-step flow with GCash/Maya integration, no third-party cut
- ✨ AI concierge (Maya) — Guest-facing chatbot for trip planning plus a Backstage co-worker mode for inbox triage
- ✨ Owner CRM — Kanban deal pipeline, dispatch board, lead tracking, and revenue analytics
- ✨ Type-safe end-to-end — Strict TypeScript with Zod schemas shared across packages, 213 passing tests
Stack: React 18 · TypeScript · Tailwind CSS · Vite · Zustand · Zod Website: lagunaya-eco-resort-website-web.vercel.app 🔒 This is a private repository. Please contact me for more details about this project.
A full-stack travel agency platform featuring an AI chat assistant, custom trip builder, booking flow, and a backstage CRM dashboard, all powered by Cerebras LLM with RAG retrieval.
- ✨ AI chat assistant — RAG-backed widget that recommends packages with prices and guides an inline booking flow
- ✨ Custom trip builder — Multi-step form with fuzzy destination autocomplete, generating downloadable PDF itineraries
- ✨ Backstage CRM — Sales pipeline Kanban, revenue KPIs, and supplier briefs auto-sent on deal close
- ✨ PWA & security — Offline support, strict CSP headers, rate limiting, and prompt injection protections
Stack: Next.js 16 · React 19 · TypeScript · Neon Postgres + pgvector · Cerebras gpt-oss-120b Website: likha-orcin.vercel.app 🔒 This is a private repository. Please contact me for more details about this project.
A multi-conversation AI chat interface powered by Google Gemini models. Users bring their own API key and chat through a modern dashboard with real-time streaming.
- ✨ Real-time streaming — SSE-powered responses with a model fallback chain (2.5 Flash → 2.0 Flash)
- ✨ Multi-conversation management — Create, switch, and delete chat threads with auto-generated titles
- ✨ Privacy-first API key handling — Bring-your-own key stored only in sessionStorage, auto-cleared on tab close
- ✨ Animated UI — Splash screen, three-column dashboard, Catppuccin dark/light themes
Stack: Next.js 16 · React 19 · TypeScript · Google Gemini API GitHub: fuwadog/Web_chat Live: web-chat-ruddy.vercel.app
A team time-tracking and timesheet platform for agencies and compliance-conscious teams, built around the append-only audit-trail engine originally proven out in Timekeeper: every entry is provable, not just logged.
- 🔨 Timesheets & approvals — submit, review, and approve team timesheets with per-project billable rates
- 🔨 Immutable audit trail — every edit preserves the prior version first; a full Original → Edited → Current history behind every entry, so "who changed this and when" is provable, not just logged
- 🔨 Project budgets & billing — track estimated vs. actual time per client/project, export timesheets to PDF/CSV/Excel for invoicing
- 🔨 Built for regulated teams — tamper-evident logging modeled on the audit-trail expectations legal, consulting, and healthcare teams already answer to
Stack: PENDING 🔒 Private repository, early development. Please contact me for more details about this project.
Deepening Expertise:
- 🏗️ Advanced System Design & Distributed Systems
- 📊 Database Query Optimization & Index Tuning
- 🐳 Kubernetes Orchestration & Production Deployments
- 🔗 Vector Databases (Pinecone, Weaviate) & RAG Systems
- 💰 Token Optimization for LLM Context Windows
- 🚀 Microservices Patterns & gRPC
Active Exploration:
- GraphQL API Design
- Event-Driven Architecture (Kafka, RabbitMQ)
- WebSocket & Real-time Protocol Optimization
- SOLID Principles in Practice
- Machine Learning Ops (MLOps)Production-First Mindset Every line of code ships with tests, documentation, and proper error handling. No shortcuts, no technical debt.
Systems Thinking I don't just build features—I design systems. Database optimization, caching strategies, scalability are built in from day one, not bolted on later.
Quick Ramp-Up Can learn new tech stacks in days and contribute meaningfully within weeks. More interested in solving problems than arguing about frameworks.
Strong Communication Clear code reviews, thorough documentation, and mentoring mindset. I believe great engineers make teams better.
Continuous Growth Actively studying distributed systems, database optimization, and advanced CS concepts. Not just chasing frameworks—building depth.
Choose your preferred channel:
Response time: ⚡ Within 24 hours