I’m Yu Fan, a tech leader with deep technical expertise and managerial vision. Formerly at Motorola, now at Mavenir, I’ve led teams for years, focusing on backend architecture and cloud-native solutions, staying abreast of AI and other frontier fields, and championing personal growth and lifelong learning.
This article demonstrates how parallel hybrid search—combining vector and keyword retrieval—enhances the reliability of agentic AI systems by delivering complete, high‑fidelity context compared with using either method alone.
This article presents a high‑reliability design pattern for agentic AI—parallel query expansion—detailing its Pydantic model, LangGraph workflow, concurrent execution with ThreadPoolExecutor, and a comparative experiment that shows improved recall and answer quality over a simple RAG pipeline.
This article explains how sharding a knowledge base and using decentralized retrieval can boost the reliability, precision, and latency of AI agents, providing code examples, a LangGraph workflow, and a performance comparison that shows a 28% speed gain over a monolithic setup.
This article introduces the redundant‑execution pattern for agentic AI, explains how running multiple identical agents in parallel can mitigate API timeouts, model crashes, and network glitches, and presents a quantitative comparison showing a jump from 60% to 80% success rate and markedly lower latency variance.
The article presents a high‑reliability design pattern for AI agents that uses parallel query expansion to generate diverse search queries, executes them concurrently with ThreadPoolExecutor, and demonstrates through a side‑by‑side RAG comparison that this approach markedly improves recall and answer quality.
This article introduces the decentralized blackboard collaboration pattern for building reliable agentic AI systems, explains its shared data space and specialist agents, demonstrates a customer‑support ticket workflow with analyzer, retriever, and draftsman agents, and highlights benefits such as error reduction, deeper specialization, and auditability.
This article introduces the Agent Assembly Line pattern—a three‑stage, parallel pipeline built with Pydantic models and ThreadPoolExecutor—to boost the throughput of AI agent systems, provides full implementation code, and demonstrates a 206% performance gain over a sequential approach.
This article walks through a reliability‑focused design pattern for agentic AI—competitive agent ensembles—by initializing diverse LLMs, defining structured Pydantic models, creating parallel competitor nodes, evaluating their outputs with a judge node, and demonstrating a 63% speedup and higher quality results.
This article presents the hierarchical agent‑group design pattern for reliable agentic AI, explains how specialized orchestrator and executor agents exchange structured Pydantic data, demonstrates a LangGraph workflow, and shows that the hierarchical approach yields faster execution (13.57 s vs 18.34 s) and higher‑quality reports compared with a monolithic agent.
This article demonstrates how a competitive ensemble of diverse LLM agents—Claude 3.5 Sonnet and two Llama 3 variants—combined with a structured evaluation node can improve both output quality and execution speed in agentic AI systems, using LangGraph, Pydantic models, and parallel execution.
