Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I'm Xuhong Yan, an AI engineer focused on building agentic workflows and tooling to automate end-to-end review, testing, documentation, and release processes. I design and implement coding-assistant style tools, retrieval-backed systems, evaluation checks, and production software integrations. I have strong experience across Python/FastAPI, TypeScript/Next.js, GitHub Actions, Docker, Neo4j, ChromaDB, and CI/CD; I specialize in reliable automation and observable systems. I thrive in engineering-agent roles that require robust orchestration, prompt testing, and secure state management across frontend and backend services.…I'm Xuhong Yan, an AI engineer focused on building agentic workflows and tooling to automate end-to-end review, testing, documentation, and release processes. I design and implement coding-assistant style tools, retrieval-backed systems, evaluation checks, and production software integrations. I have strong experience across Python/FastAPI, TypeScript/Next.js, GitHub Actions, Docker, Neo4j, ChromaDB, and CI/CD; I specialize in reliable automation and observable systems. I thrive in engineering-agent roles that require robust orchestration, prompt testing, and secure state management across frontend and backend services.

Xuhong Yan

AI Engineer, Full Stack Developer, Game Developer, +2





I'm Xuhong Yan, an AI engineer focused on building agentic workflows and tooling to automate end-to-end review, testing, documentation, and release processes. I design and implement coding-assistant style tools, retrieval-backed systems, evaluation checks, and production software integrations. I have strong experience across Python/FastAPI, TypeScript/Next.js, GitHub Actions, Docker, Neo4j, ChromaDB, and CI/CD; I specialize in reliable automation and observable systems. I thrive in engineering-agent roles that require robust orchestration, prompt testing, and secure state management across frontend and backend services.…I'm Xuhong Yan, an AI engineer focused on building agentic workflows and tooling to automate end-to-end review, testing, documentation, and release processes. I design and implement coding-assistant style tools, retrieval-backed systems, evaluation checks, and production software integrations. I have strong experience across Python/FastAPI, TypeScript/Next.js, GitHub Actions, Docker, Neo4j, ChromaDB, and CI/CD; I specialize in reliable automation and observable systems. I thrive in engineering-agent roles that require robust orchestration, prompt testing, and secure state management across frontend and backend services.

Available to hire

I’m Xuhong Yan, an AI engineer focused on building agentic workflows and tooling to automate end-to-end review, testing, documentation, and release processes. I design and implement coding-assistant style tools, retrieval-backed systems, evaluation checks, and production software integrations.

I have strong experience across Python/FastAPI, TypeScript/Next.js, GitHub Actions, Docker, Neo4j, ChromaDB, and CI/CD; I specialize in reliable automation and observable systems. I thrive in engineering-agent roles that require robust orchestration, prompt testing, and secure state management across frontend and backend services.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Intermediate

Language

English

Fluent

Work Experience

Full Stack AI Developer at Aicyber

December 1, 2025 - Present

Built and maintained production agent workflows for a live AI legal information assistant, including case analysis, legal Q&A, document generation, evidence summary, brief generation, and contract review. Implemented LLM orchestration with LangGraph, CopilotKit, OpenAI, Anthropic, retrieval, tool calling, persistent conversation state, structured outputs, and LangSmith observability. Improved reliability through prompt test cases, RAG answer-quality checks, history hydration, clean thread switching, and robust state handling across frontend and backend services. Developed authenticated product flows across FastAPI, Next.js, Supabase auth/database, DigitalOcean deployment, Vercel frontend deployment, and protected chat/account routes. Built SP400 quantitative AI infrastructure spanning data pipelines, LLM factor discovery, backtesting, validation, forward testing, monitoring, execution support, and a simple dashboard.

Intern at AI Agents Platform

June 30, 2025 - July 23, 2025

Developed a comprehensive AI digital human management platform with secure JWT authentication. Built scalable RESTful API backend using FastAPI with professional error handling. Created responsive frontend applications using Next.js App Router and TypeScript. Designed admin dashboard with analytics and user management capabilities. Implemented Docker containerization for streamlined development environment. Utilized type-safe development and microservices architecture with proper ORM relationships and migration support.

Full Stack Developer Intern at Soontech

June 1, 2025 - October 1, 2025

Built a customizable conversational AI platform using LangChain and LangGraph with training and chatting modes, intent recognition, memory extraction, and context-aware response generation. Developed a FastAPI backend with 15+ REST endpoints, JWT authentication, Pydantic validation, and SSE streaming for real-time AI responses. Implemented a memory pipeline that extracted keywords and relationships from conversations, storing graph relationships in Neo4j and semantic memories in ChromaDB. Engineered a Next.js 15 and TypeScript frontend with protected routes, token management, reusable API clients, and authenticated flows across 8+ feature pages. Containerised MySQL, MongoDB, Neo4j, and ChromaDB services with Docker Compose for repeatable local setup and integration testing.

Intern at AI Agents Platform

June 1, 2025 - Present

Developed a comprehensive AI digital human management platform featuring secure JWT authentication. Built a scalable RESTful API backend using FastAPI with professional error handling. Created responsive frontend applications with Next.js App Router and TypeScript. Designed admin dashboard with analytics and user management features. Implemented Docker containerization for a streamlined development environment. Focused on type-safe development using TypeScript and Pydantic schemas, a microservices architecture with separation of concerns, and database schema design with proper ORM relationships and migration support.

Student (Project) at University of Melbourne

November 30, 2024 - July 23, 2025

Developed hybrid PDDL planning and Q-learning agent for maze navigation. Designed PDDL domain/problem files for strategic high-level planning. Implemented Q-learning with custom feature extraction and parameter optimization. Reduced repetitive movement patterns by 90% through position-history loop detection. Conducted ablation studies demonstrating 25-35% performance impact of key features.

AI Course Project Developer at University of Melbourne

November 30, 2024 - July 23, 2025

Developed a hybrid AI agent combining PDDL planning and Q-learning for maze navigation. Designed PDDL domain and problem files for high-level strategic planning. Implemented Q-learning algorithms with custom feature extraction and parameter optimization. Significantly reduced repetitive movement patterns by 90% using position-history loop detection and conducted ablation studies demonstrating 25-35% performance impact of key features.

Team Member at Multi-University Robotics Competition

July 31, 2021 - July 23, 2025

Led mechanical design and load-bearing capacity calculations. Created 3D models using SolidWorks and programmed operational logic with AD2. Managed fabrication process including 3D printing and assembly. Successfully completed precision ball placement task in approximately 18 seconds, demonstrating efficient design.

Team Leader & Mechanical Designer at Multi-University Robotics Competition

July 31, 2021 - July 23, 2025

Led mechanical design and load-bearing calculations for an autonomous robot tasked with precision ball placement. Created 3D models using SolidWorks and programmed operational logic with AD2. Managed the fabrication process including 3D printing and assembly of the robot. Successfully completed the task in approximately 18 seconds, demonstrating effective and efficient design.