Author

Chengwu Tech Stack

A powerful mindset is a lifelong treasure!

Articles

Likes

Views

Comments

Latest from Chengwu Tech Stack

37 recent articles

Chengwu Tech Stack

Oct 8, 2026 · Artificial Intelligence

RAG from Prototype to Production: Scenarios, Pitfalls, and Deployment Strategies

This article explains Retrieval-Augmented Generation (RAG) fundamentals, identifies suitable use cases, details six common production pitfalls like document parsing errors and permission leaks, outlines required modules for production RAG systems, and describes evaluation and incremental rollout strategies for reliable deployment.

Chunking StrategiesDocument ParsingHybrid Search

0 likes · 25 min read

RAG from Prototype to Production: Scenarios, Pitfalls, and Deployment Strategies

Chengwu Tech Stack

Sep 8, 2026 · Artificial Intelligence

From Video Analysis Pipeline to Production System: Architecture, Contracts, and Guardrails

This article details the system architecture for production-ready video AI analysis, covering task contracts with three distinct identifiers, media and model isolation, capacity estimation using Little's Law, retry and recovery rules with outbox pattern, evidence retention requirements, and a phased gray-release strategy for two scenarios: engineering quality inspection and intelligent security.

capacity planningevidence retentiongray release

0 likes · 16 min read

From Video Analysis Pipeline to Production System: Architecture, Contracts, and Guardrails

Chengwu Tech Stack

Sep 7, 2026 · Artificial Intelligence

LangChain Video Analysis: Fixed Workflows First, Bounded Agents Only When Needed

This tutorial demonstrates building a production-ready video analysis pipeline using LangChain and LangGraph, emphasizing a fixed five-node workflow with structured state management, and only introducing a constrained agent for dynamic evidence gathering when fixed rules are insufficient, plus failure recovery and acceptance testing strategies.

Agent OrchestrationLangChainLangGraph

0 likes · 15 min read

LangChain Video Analysis: Fixed Workflows First, Bounded Agents Only When Needed

Chengwu Tech Stack

Sep 6, 2026 · Artificial Intelligence

Building an Explainable Person-Detection Contract for Video Frame Sampling

This tutorial defines a three-state judgment contract (present, not_observed, uncertain) for detecting real persons in sampled video frames, covering prompt design with structured JSON output, evidence requirements, code-level aggregation rules, and evaluation methodology using boundary-case samples.

AI visionevaluation methodologyjudgment contract

0 likes · 15 min read

Building an Explainable Person-Detection Contract for Video Frame Sampling

Chengwu Tech Stack

Sep 4, 2026 · Artificial Intelligence

Will AI Get Smarter with Use? Building a Verifiable Feedback-to-Improvement Pipeline

This article outlines a rigorous engineering pipeline for turning human feedback into verified AI improvements, covering fact verification, six-layer root-cause diagnosis, experiment cards with falsifiable hypotheses, segregated test sets to prevent data leakage, and controlled releases with rollback plans, illustrated via video inspection and security alert case studies.

AI engineeringcontinuous improvementcontrolled release

0 likes · 29 min read

Will AI Get Smarter with Use? Building a Verifiable Feedback-to-Improvement Pipeline

Chengwu Tech Stack

Sep 3, 2026 · Artificial Intelligence

Designing a Resilient AI Runtime: Observability, Degradation, Takeover, Rollback

This article outlines a production-grade AI runtime architecture that separates processing state from business conclusions, enforces bounded model execution with retry budgets and circuit breakers, ensures idempotent external actions, defines explicit degradation paths, structures human-in-the-loop workflows, extends observability to judgment quality, and validates rollback and failure injection before launch.

AI productioncircuit breakerdegradation

0 likes · 27 min read

Designing a Resilient AI Runtime: Observability, Degradation, Takeover, Rollback

Chengwu Tech Stack

Sep 2, 2026 · Artificial Intelligence

Beyond Accuracy: A Four-Layer Framework for Production-Ready AI Evaluation

The article argues that model accuracy alone is insufficient for AI production deployment and proposes a four-layer evaluation framework covering model capability, task contract, system engineering, and business results, along with five balanced metric groups, risk-based test datasets, staged deployment gates, and continuous regression testing.

AI engineeringAI evaluationdeployment gates

0 likes · 27 min read

Beyond Accuracy: A Four-Layer Framework for Production-Ready AI Evaluation

Chengwu Tech Stack

Sep 1, 2026 · Artificial Intelligence

Agent vs Prompt+LLM: Where Should Uncertainty Live in Production LLM Systems?

This article analyzes when to use autonomous agents versus prompt-engineered LLM pipelines in production systems, arguing that the choice depends on uncertainty type, failure cost, reversibility, and audit requirements, and presents a layered architecture where code handles deterministic boundaries while LLMs handle semantic judgments.

AI system designAgent vs Prompt engineeringLLM architecture

0 likes · 25 min read

Agent vs Prompt+LLM: Where Should Uncertainty Live in Production LLM Systems?

Chengwu Tech Stack

Aug 31, 2026 · Artificial Intelligence

Autonomous vs. Engineered Agents: Balancing Flexibility and Control in Production

This article distinguishes autonomous agents that dynamically plan actions from engineered agents that embed decision-making within verifiable, auditable workflows, providing criteria for choosing each approach, a three-layer production architecture, and a checklist for determining the right level of autonomy based on risk, reversibility, and audit requirements.

AI agentsAgent Architectureagent evaluation

0 likes · 16 min read

Autonomous vs. Engineered Agents: Balancing Flexibility and Control in Production

Chengwu Tech Stack

Aug 21, 2026 · Artificial Intelligence

Reasoning Models vs World Models: Why AI Needs Both for Autonomous Action

This article distinguishes reasoning models from world models, arguing that personal AI agents require external, continuously updated world models to track current state, predict action consequences, and learn from prediction errors, proposing a five-layer hybrid architecture and phased implementation roadmap.

AI architectureautonomous agentspersonal AI

0 likes · 25 min read

Reasoning Models vs World Models: Why AI Needs Both for Autonomous Action