Jeff Xu

Background

About Jeff Xu

I build production AI systems that solve real business problems. Currently: Architect & solo developer of TingleMind, a dual-system AI platform for financial analysis combining conversational multi-agent orchestration with event-driven ML forecasting. Technical Achievements: • 83,505 lines of production Python across 14 specialized AI agent groups • Trained GMM clustering model for 45K+ economic news events (99.4% coverage) • Built Silent Listener with 7 layers of intelligence (patent-worthy) • Integrated GPT-4/5, Pinecone vector DB, MongoDB Atlas, FRED data • Designed exponential decay kernels for temporal impact attribution • Deployed to AWS with 99%+ uptime serving real-time analysis Core Expertise: • Multi-agent orchestration with LLM routing • Unsupervised machine learning (GMM clustering) • Semantic search using embedding transformers (800K+ vectors) • Event-driven forecasting with causal attribution • Distributed caching & async job scheduling • Production deployment (FastAPI, MongoDB, AWS) I'm passionate about building AI-native products that are transparent, explainable, and genuinely useful.