I'm @Stan370, a full-stack engineer and entrepreneur with startup and remote-team experience. I ship products end to end — React/TypeScript frontends, Go/Python/Node backends, edge deployments — and I'm fluent in AI-assisted development. I think in terms of users and business outcomes: one of my games has recorded ~8,500 game sessions from ~1,000 players.
A data-driven web application that analyzes League of Legends players' MBTI personality types based on champion pools, playstyles, and in-game behavior patterns.
A new version of an old tile-match UGC game, incorporating features like daily challenges, a Reddit scheduler, and an auto-post bot, built using Devvit 0.12.
An interactive puzzle game with real-time leaderboards and multi-language support. Implemented complex grid logic and collision detection in TypeScript. Integrated Redis for optimized leaderboard management and resolved ESM/CommonJS compatibility issues for Devvit/Webview integration.
An authentication sidecar designed for zero-trust Kubernetes environments. Handles token introspection and RBAC enforcement at the pod level.
Event-driven architecture for managing stateful LLM agents. Features a custom vector retrieval pipeline and optimistic concurrency control for agent memory.
Designed and built a multi-tenant agent marketplace enabling users to deploy custom AI assistants with private knowledge. Supports dynamic model switching across OpenAI, Claude, and Gemini. Features secure user authentication and real-time chat history management.
An autonomous agent interface for managing Valorant Champions Tour fantasy leagues. Scrapes real-time data and proposes roster changes.
An AI-powered platform designed to help teachers handle multilingual classrooms effortlessly. It turns spoken lessons, story reading, or student discussions into real-time text, automatically translating them into multiple languages so every student can follow along.