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

Block or report ezhanjaved

Report abuse

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

Report abuse
ezhanjaved/README.md

Hi, I’m Ezhan πŸ‘‹

Full Stack Software Engineer building scalable SaaS products, with a strong focus on systems, automation, and applied AI.

I enjoy tackling real problems through system design and automation, and experimenting with reinforcement learning in simulated environments, with a strong interest in how reinforcement learning and game theory intersect.


What I’m Working On?

Locanize β€” Local SEO SaaS (Founder & Engineer)

πŸ”— https://www.locanize.com

A full-scale Local SEO platform built to automate audits, tracking, and competitive insights for businesses.

Highlights

  • Built using React, Node, and FastAPI
  • Designed async pipelines with Celery & RabbitMQ
  • Automated grid tracking, rank volatility, and competitor analysis
  • Scaled to 100+ businesses by removing manual workflows

RL Playground β€” Reinforcement Learning Sandbox

An experimental environment to explore agent training in simulated worlds.

Focus Areas

  • 3D environments with React Three Fiber & Rapier
  • PPO-based training pipeline using FastAPI & PyBullet
  • Reward shaping, environment resets, and benchmarking
  • Stable-Baselines3 for policy optimization

ZEDGE β€” 2D AI-Driven Web Game (Live)

πŸ”— Live: https://ai-frontend-theta.vercel.app/

A deployed 2D browser-based game built with Phaser.js, integrating reinforcement learning agents for in-game decision making.

Highlights

  • Real-time gameplay mechanics using Phaser.js
  • FastAPI backend managing environment state and agent training
  • Implemented reinforcement learning agents using Q-Learning and PPO
  • Fine-tuned GPT-2 to generate dynamic trivia and in-game content

πŸ›  Tech Stack

Languages
JavaScript, TypeScript, Python, PHP, SQL

Backend
FastAPI, Node.js, Express, Celery, RabbitMQ, Redis

Frontend
React, React Three Fiber, Phaser.js

AI / ML
LLMs (GPT-2, RAG), Reinforcement Learning (Q-Learning, PPO, DQN, TRPO)

Databases
PostgreSQL, MySQL, MongoDB

Other
SaaS Architecture, Web Automation, SEO Systems, DevOps


✍️ Writing & Articles

I write about software, AI, game engines, optimization, tech culture, and more.

πŸ“– Featured Articles


πŸ“š Background

  • BSc Software Engineering β€” UIT University
  • Founder & Full Stack Engineer β€” Locanize
  • Previously worked across web development, SEO systems, and eCommerce platforms

πŸ“« Connect

Pinned Loading

  1. rl-playground rl-playground Public

    A Reinforcement Learning Playground built on top of Three.js + Rapier.js + PyBullet

    JavaScript 2

  2. zedge-backend zedge-backend Public

    Backend API for delegating real-time responses to game AI using the trained Reinforcement Learning models (Q-Learning & PPO) as well GPT-2 for trivia based rounds

    Python

  3. zedge-frontend zedge-frontend Public

    A web-based game that is developed using JavaScript gaming engine Phaser.js

    JavaScript