╭─────────────────────────────────────────────╮
    │  Co-Founder @ Brixsport                    │
    │  Full-Stack • ML Engineering • Systems     │
    │  📍 Lagos, Nigeria                         │
    ╰─────────────────────────────────────────────╯

Currently shipping Brixsport — a real-time sports analytics platform serving university tournaments across Nigeria. We're solving the problem of fragmented campus sports data through event-driven architecture and progressive web apps designed for low-bandwidth environments.

Technical challenges solved:

• Sub-second live match updates with Redis pub/sub
• PWA-based stat recording that works on 2G networks
• Scalable database design handling concurrent tournaments
• Event sourcing patterns for match state management

Stack: Next.js • Supabase • Redis/Upstash • TailwindCSS

const expertise = {
  backend: {
    focus: "Event-driven systems, API design, database architecture",
    tools: ["Node.js", "PostgreSQL", "Redis", "Supabase"],
    strength: "Building infrastructure that scales under real load"
  },
  
  ml_data: {
    focus: "Feature engineering, geospatial analysis, production ML",
    tools: ["Python", "scikit-learn", "XGBoost"],
    strength: "Models that don't leak data or overfit in production"
  },
  
  frontend: {
    focus: "React ecosystem, performance optimization, PWAs",
    tools: ["React", "Next.js", "TailwindCSS"],
    strength: "Interfaces that work in unreliable network conditions"
  }
}

Built an XGBoost-based delivery estimation model achieving ~77% relevance through proper temporal validation and engineered geospatial features. Key focus: eliminating data leakage and proxy variables that inflate validation scores.

Methodology: 70/15/15 temporal split • Haversine distance features • Hyperparameter grid search

Real-time fault detection and quarantine system for multi-core task execution. Implemented health monitoring logic with live visualization using Python, Tkinter, and Matplotlib.

Key features: Task scheduler • Fault detection algorithm • Dynamic core quarantine • Real-time metrics dashboard

Philosophy: Ship code that survives production, not just demos. I care about:

• ✅ Temporal validation in ML (not just random splits)
• ✅ Database schemas that handle real concurrency
• ✅ APIs that fail gracefully under load
• ✅ Features that work on slow networks
• ❌ Over-engineering simple problems
• ❌ Optimizing for metrics that don't matter

Current learning: Distributed systems patterns, advanced Postgres optimization, WebSocket scaling strategies

I'm interested in collaborating on:

• Backend infrastructure projects that need to scale
• ML engineering work with real production constraints
• Real-time systems where latency and reliability matter

Best way to reach me: richardogunwole17@gmail.com