Game Engine Knowledge Base
Shows the ability to turn fragmented domain knowledge into governed data assets and AI-ready content systems.
Approved U.S. immigrant visaOpen to relocation in the U.S.
Open to U.S. AI engineering roles
Theo Yang
AI Engineer
Enterprise AI Applications · Data & Risk
AI engineer with a 16-year career spanning data engineering, machine learning and risk systems, enterprise data platforms, and AI applications. I build from problem discovery and solution design through production deployment, enterprise integration, and user adoption.
Current focus
Enterprise AI applications
Production proof
15M API calls/day
Domain edge
Data & risk systems
Engineering Journey
Data Engineering→Data Mining & ML→Risk & Data Platforms→AI Applications→Forward-Deployed Delivery
My career has progressed from ETL and large-scale analytics to recommendation, credit intelligence, logistics analytics, ChatBI, and production AI agents. That path lets me connect model behavior with data foundations, domain constraints, APIs, deployment, and measurable adoption.
Target Role Families
I am focused on U.S. roles where hands-on AI application engineering meets enterprise delivery and domain depth:
Primary Entry
Lead- AI Engineer
- AI Application Engineer
- Applied AI Engineer
Forward-Deployed Fit
- Forward Deployed Engineer
- AI Solutions Engineer
- Enterprise AI Engineer
Domain Advantage
- Risk AI Engineer
- Fraud AI Engineer
- Data & AI Engineer
Core Capabilities
1
LLM Application Engineering
Building RAG pipelines, prompt and instruction workflows, LoRA-based adaptation, and grounded enterprise AI applications.
2
Agentic Systems
Designing ReAct-style tool orchestration, multi-step analytical workflows, fallback paths, and human-in-the-loop controls.
3
Evaluation & Reliability
Evaluating retrieval, generation, hallucination risk, business outcomes, model drift, and production behavior.
4
Data & Machine Learning
Working across Python, SQL, ETL, feature engineering, XGBoost, FAISS, Hadoop, and Spark MLlib.
5
Production & Enterprise Delivery
Taking systems from problem definition to APIs, secure deployment, integration, monitoring, scale, and user adoption.
AI Engineering Lab
A focused set of practical systems showing how I turn messy domain knowledge, files, videos, and workflows into searchable, testable, and reusable AI-assisted tools.
Game Config & Lua Analysis Toolkit
Shows practical engineering of operational rules, validation logic, and domain-specific developer tooling.
Video-to-Documentation Pipeline
Shows the ability to convert unstructured media into reusable learning assets with measurable production efficiency.
AI Coding Workflow Playbook
Shows that I can teach and operationalize AI-assisted development instead of only using tools ad hoc.
Selected AI Systems
Fraud AILLM + RAGLoRASecure DeploymentHuman-in-the-loop
Enterprise Fraud Detection AI Platform
Built a production fraud-detection agent for regulated telecom and public-sector scenarios, combining deterministic rules, NLP, LLM reasoning, RAG, secure deployment, and standardized APIs at 15M calls per day.
Proof signal
Demonstrates production AI engineering under real security, latency, integration, scale, and adoption constraints—not a standalone model demo.
NL2SQLTiered RoutingReAct AgentFAISSEnterprise Analytics
Enterprise ChatBI / NL2SQL System
Built a ChatBI system for logistics analytics that replaced an unreliable pure-LLM path with tiered routing, guided clarification, retrieval, validation, and ReAct-style tool orchestration.
Proof signal
Shows the ability to engineer reliable enterprise LLM applications by combining probabilistic reasoning with deterministic routing, tools, data context, and validation.
Production MLXGBoostFeature EngineeringRisk AIAPI Platform
Xiaomi Credit Intelligence & Risk Platform
Built a credit-scoring platform from scratch for thin-file users, converting MIUI, IoT-device, and e-commerce behavior into 1,500 features, an XGBoost model, and API-based risk services.
Proof signal
Demonstrates end-to-end production ML engineering: feature systems, interpretable modeling, APIs, scale, partner integration, and measurable financial impact.
Real-time AnalyticsRoot-cause AnalysisLogisticsDecision Support
Meituan Logistics Early-Warning & Diagnosis System
Built a real-time early-warning and diagnosis system for logistics fulfillment, connecting operational signals to root-cause analysis and measurable improvement actions.
Proof signal
Shows how large-scale data systems become operational engineering tools that improve reliability, cost, and frontline decision quality.