Open to opportunities

gao bin

@gaobin1122

Message

I build high-concurrency Golang backends for real-time games and AI systems.

China

Message

What I'm looking for

I'm looking for a Golang backend role where I can build high-concurrency, high-availability distributed services, own architecture through production operations, and apply LLM and RAG capabilities to intelligent game or backend systems.

I've led backend delivery for SimDunk 3v3, building distributed login, gateway, combat, and matchmaking services that scale for low-latency real-time play. I improved matchmaking success by 35%, reduced average queue time by 50%, and cut peak server CPU consumption by 40% through GC tuning, object pooling, packet compression, and SQL governance.

Across Little Shop and A3A, I've built microservice architectures, game systems, Redis-backed leaderboards, payment and anti-cheat integrations, and production operations workflows. I also build AI-driven game and agent systems with Golang, Python, LangChain, LangGraph, RAG, vector databases, and LLM APIs.

Experience

Work history, roles, and key accomplishments

Current

Independent Development

Current

A-Share Intelligent Investment Analysis Agent

Apr 2026 - Present (4 months)

Built 6-stage LLM inference pipeline based on LangGraph to realize natural language intent recognition, multi-source data collection, automated stock screening and structured report generation. Designed three-tier data source degradation + TTL cache strategy to reduce LLM hallucinations; real-time streaming SSE push of inference process.

Python LangGraph fastAPI SSE Streaming Push LLM LLM Inference Data Collection Report Generation Data Source Degradation TTL Cache Streaming

Current

Full-Stack Independent Development

Current

AI-Driven Interactive Story Game System

Feb 2026 - Present (6 months)

Layered architecture design: Python pipeline for RAG novel knowledge extraction (character graphs, events, semantic vectors); Golang real-time game server with interoperable CSV configuration to lightweightly integrate AI capabilities into game backend. Self-developed WebSocket long connection engine on Go server, supporting AI character dialogue, scene teleportation, LLM-powered combat and asynch

Python LangGraph ChromaDB (Vector Database)WebSockets MySQL Redis RAG Vector Databases AI Integration Vendor Agnostic LLM Adaptation Asynchronous LLM Calls

Golang Backend Lead

A3A Shooting & Progression Mini Game

Aug 2024 - Jun 2026 (1 year 10 months)

Independently built server framework and project layered structure; led development of core systems including shooting combat, equipment, skills and leaderboards. Adopted separated login & business multi-process architecture with efficient inter-process communication. Redis cluster hosts cross-server leaderboards and temporary event data to guarantee stable performance under mini-game peak concurr

Multi Process Redis Cluster Protobuf Telegram SDK Cross Server Leaderboards Peak Concurrency Payment Integration Anti Cheat Logic Pressure Testing

Golang Backend Engineer

SimDunk

Dec 2022 - Jun 2026 (3 years 6 months)

Built distributed game service architecture from scratch with separated login, gateway, combat and matchmaking servers, enabling horizontal scaling to support thousands of concurrent players in low-latency real-time battles. Refactored matchmaking scheduling logic, optimized player pool and rank-based matching algorithms: matchmaking success rate up 35%, average queue time reduced by 50%.

Self Built RPC Protobuf Redis Cluster MySQL MongoDB Linux Distributed Systems Performance Optimization Matchmaking GC Tuning Object Pool Slow SQL Governance Grayscale Releases

Golang Backend Developer

Little Shop

Aug 2020 - Dec 2022 (2 years 4 months)

Designed multi-process distributed microservice architecture with self-defined RPC inter-service communication, supporting elastic scaling to handle sudden peak traffic during operational activities. Designed and developed full business modules including alliance, shop, quests and events; independently refactored distributed mail service with 70% throughput improvement.