AI in production

7,000 RPS at peaks

$20M+ TVL
10K+ users

Profile

AI-first fullstack developer building high-performance systems in Web3, fintech, and beyond.

From developer to leader.
From zero to production.

Capabilities

Build from zero

Design and develop products from the ground up with clear architecture and intent.

Scale production

Optimize performance, reliability, and cost under real production load.

Ship with AI

Use agents and applied AI to accelerate delivery without giving up engineering review.

TechStack

Go, Python, PHP, TypeScript, React, PostgreSQL, Redis, Docker, Linux, CI, Claude Code, MCP, and Solidity.

Job experience

orion/lumia, Lead Fullstack Developer

Go / React / Redis / RWA

  • throughput: 7,000 RPS at peaks / 99.9% uptime
  • cache-efficiency: database reads reduced by 79%
  • order-latency: 300 ms to 50 ms
  • delivery-pipeline: AI-assisted branch-to-PR workflow
  • cross-chain: 1inch, Uniswap, PancakeSwap integrated

UFarm.Digital, Lead Solidity / Fullstack Developer

Python / FastAPI / React / Solidity

  • platform-scale: 10K+ users / $20M+ TVL
  • quality-coverage: 45% to 85%
  • feature-delivery: 2 days to 1 hour
  • onboarding: 2 weeks to 3 days
  • infrastructure: operating costs reduced by 35%
  • leadership: 4 developers / 95% sprint delivery

Rock And Block, Fullstack Developer

PHP / Laravel / React / Vue.js

  • client-delivery: 12+ projects / every deadline met
  • test-runtime: execution time reduced by 30%
  • fullstack: Laravel APIs / React and Vue products
  • mentoring: 3 interns / 2 permanent hires
  • web3-delivery: DeFi, NFT, token launches

DINS, System Administrator / Python Developer

Python / Django / Pandas / networks

  • reporting: manual effort reduced by 80%
  • time-saved: about 15 hours per week
  • automation: Django and Pandas reporting system
  • infrastructure: 200+ office workstations
  • reliability: SLA compliance maintained

Selected work

pipi

An isolated coding harness for serious tasks, powered by Pi.

View repository ↗

  1. One environment

  2. Specialized agents

  3. Audited delivery

  4. Live steering

Search, implementation, browser debugging, reviews, and long-running commands stay inside one focused workspace.

pipi://workspace
The main Pipi workspace with its command input and project status.

My agent skills

Focused agent capabilities built as small, inspectable packages for real engineering work.