# Data Leads Future
> One practical story every month, sharing my hard-learned experiences in the enterprise AI space. Every article saves you 40+ hours of work.
Public Ghost content for AI and LLM tooling. Use `/llms-full.txt` for consolidated page and post context.
Append `.md` to any post or page URL to get the content in Markdown (for example, `/example-post.md`).

## Pages
- [About this site](https://www.dataleadsfuture.com/about.md) - About me Who am I Hi, I'm Peng Qian. I was a senior data scientist at Alibaba. Now I'm the chief architect of the big data department at China Merchants Securities, one of China's largest investment banks. I lead research on the integration of data and AI. What I do: I don't want data or AI to be t…
- [Contact](https://www.dataleadsfuture.com/contact.md) - You can find me on: Twitter, LinkedIn, Facebook. If you have any questions about Data Leads Future, you can contact me at dataleadsfuture@gmail.com.
- [Cookies](https://www.dataleadsfuture.com/cookies.md) - COOKIE POLICY Last updated August 27, 2023 This Cookie Policy explains how Data Leads Future ("Company," "we," "us," and "our") uses cookies and similar technologies to recognize you when you visit our website at https://www.dataleadsfuture.com ("Website"). It explains what these technologies are a…
- [Disclaimer for Data Leads Future](https://www.dataleadsfuture.com/disclaimer.md) - Please read this disclaimer carefully before using the Data Leads Future website (the "Website") operated by Peng Qian ("us", "we", "our"). General Information Purposes Only The information provided by Data Leads Future is for general informational purposes only. All information on the website is p…
- [Privacy Policy](https://www.dataleadsfuture.com/privacy-policy.md) - PRIVACY POLICY Last updated August 27, 2023 This privacy notice for Data Leads Future ("we," "us," or "our"), describes how and why we might collect, store, use, and/or share ("process") your information when you use our services ("Services"), such as when you: * Visit our website at https://www.da…
- [Tags](https://www.dataleadsfuture.com/tags.md) - Harness Engineering - Data Leads FutureOne practical story every month, sharing my hard-learned experiences in the enterprise AI space. Every article saves you 40+ hours of work.Data Leads FuturePeng QianAgentic AI - Data Leads FutureStay ahead in the world of AI and data science! Whether you’re a…
- [Terms and Services for Data Leads Future](https://www.dataleadsfuture.com/terms-and-services.md) - Introduction Welcome to Data Leads Future (https://www.dataleadsfuture.com), a blog dedicated to providing educational content on data science, including tutorials, practical applications, coding techniques, machine learning, and algorithmic trading in stock and currency markets. This Terms and Ser…

## Posts
- [No Plugins Needed, I Built a Fully Automated Coding Loop in OpenCode](https://www.dataleadsfuture.com/no-plugins-needed-i-built-a-fully-automated-coding-loop-in-opencode.md) - Using DeepSeek-V4 for low-cost Loop Engineering
- [Why Well-Written Prompts Are Almost Always Structured](https://www.dataleadsfuture.com/why-well-written-prompts-are-almost-always-structured.md) - A Plain and Intuitive Explanation
- [DeepSeek-V4 Can't Read Images? I Made It Read](https://www.dataleadsfuture.com/deepseek-v4-cant-read-images-i-made-it-read.md) - Don't wait for a multimodal model, you can use it now
- [Reflection SDD: Use a Reflection Harness to Level Up Your OpenSpec Workflow](https://www.dataleadsfuture.com/reflection-sdd-use-a-reflection-harness-to-level-up-your-openspec-workflow.md) - Stop letting bad spec files tank your code quality
- [How I Use OpenCode, Oh-My-OpenCode-Slim, and OpenSpec to Build My Own AI Coding Environment](https://www.dataleadsfuture.com/how-i-use-opencode-oh-my-opencode-slim-and-openspec-to-build-my-own-ai-coding-environment.md) - Ride the wave of AI coding, don't get swept away by it
- [How to Use Agent Skills in Enterprise LLM Agent Systems](https://www.dataleadsfuture.com/how-to-use-agent-skills-in-enterprise-llm-agent-systems.md) - A thorough and detailed hands-on guide
- [Advanced RedisVL Long-term Memory Tutorial: Using an LLM to Extract Memories](https://www.dataleadsfuture.com/advanced-redisvl-long-term-memory-tutorial-using-an-llm-to-extract-memories.md) - Building an intelligent, context-aware memory system
- [Build Long-Term and Short-Term Memory for Agents Using RedisVL](https://www.dataleadsfuture.com/build-long-term-and-short-term-memory-for-agents-using-redisvl.md) - Pros and cons analysis based on real-world practice
- [Microsoft Agent Framework (MAF) Middleware Basics: Add Compliance Fences to Your Agent](https://www.dataleadsfuture.com/microsoft-agent-framework-maf-middleware-basics-add-compliance-fences-to-your-agent.md) - Principles, usage, and development standards
- [My Agent System Looks Powerful but Is Just Industrial Trash](https://www.dataleadsfuture.com/my-agent-system-looks-powerful-but-is-just-industrial-trash.md) - Single agent, low robustness, and position bias
- [Make Microsoft Agent Framework’s Structured Output Work With Qwen and DeepSeek Models](https://www.dataleadsfuture.com/make-microsoft-agent-frameworks-structured-output-work-with-qwen-and-deepseek-models.md) - Updated to the latest version of Microsoft Agent Framework
- [A Quick Guide to Containerizing Agent Applications with Podman](https://www.dataleadsfuture.com/a-quick-guide-to-containerizing-agent-applications-with-podman.md) - Alternative solutions compatible with Docker SDK
- [Share My LLM Prompts and Tips That Make Work and Learning Super Efficient](https://www.dataleadsfuture.com/share-my-llm-prompts-and-tips-that-make-work-and-learning-super-efficient.md) - Can they help you too?
- [Exclusive Reveal: Code Sandbox Tech Behind Manus and Claude Agent Skills](https://www.dataleadsfuture.com/exclusive-reveal-code-sandbox-tech-behind-manus-and-claude-agent-skills.md) - Use Jupyter code executor to help your agent finish tasks in a smarter way
- [How To Use DeepSeek-OCR And Docling For PDF Parsing](https://www.dataleadsfuture.com/how-to-use-deepseek-ocr-and-docling-for-pdf-parsing.md) - Is it really that good
- [How I Crushed Advent of Code And Solved Hard Problems Using Autogen Jupyter Executor and Qwen3](https://www.dataleadsfuture.com/how-i-crushed-advent-of-code-and-solved-hard-problems-using-autogen-jupyter-executor-and-qwen3.md) - A detailed guide on using AI to generate code and solve puzzles automatically and quickly
- [I Used Autogen GraphFlow and Qwen3 Coder to Solve Math Problems — And It Worked](https://www.dataleadsfuture.com/i-used-autogen-graphflow-and-qwen3-coder-to-solve-math-problems-and-it-worked.md) - More reliable than your math professor
- [Use LLamaIndex Workflow to Create an Ink Painting Style Image Generation Workflow](https://www.dataleadsfuture.com/use-llamaindex-workflow-to-create-an-ink-painting-style-image-generation-workflow.md) - Add strong artistic flair through fine control of LLM context
- [Monitoring Qwen 3 Agents with MLflow 3.x: End-to-End Tracing Tutorial](https://www.dataleadsfuture.com/monitoring-qwen-3-agents-with-mlflow-3-x-end-to-end-tracking-tutorial.md) - Enhance your multi-agent application's observability, explainability and Traceability
- [Build AutoGen Agents with Qwen3: Structured Output & Thinking Mode](https://www.dataleadsfuture.com/build-autogen-agents-with-qwen3-structured-output-thinking-mode.md) - Save yourself 40 hours of trial and error
- [Fixing the Agent Handoff Problem in LlamaIndex's AgentWorkflow System](https://www.dataleadsfuture.com/fixing-the-agent-handoff-problem-in-llamaindexs-agentworkflow-system.md) - The position bias in LLMs is the root cause of the problem
- [Integrating LlamaIndex and DeepSeek-R1 for reasoning_content and Function Calling Features](https://www.dataleadsfuture.com/integrating-llamaindex-and-deepseek-r1-for-reasoning_content-and-function-call-features-2.md) - Empowering AgentWorkflow with the strong boost from DeepSeek-R1
- [Diving into LlamaIndex AgentWorkflow: A Nearly Perfect Multi-Agent Orchestration Solution](https://www.dataleadsfuture.com/diving-into-llamaindex-agentworkflow-a-nearly-perfect-multi-agent-orchestration-solution.md) - And fix the issue where the agent can't continue with past requests
- [Using LLamaIndex Workflow to Implement an Agent Handoff Feature Like OpenAI Swarm](https://www.dataleadsfuture.com/using-llamaindex-workflow-to-implement-an-agent-handoff-feature-like-openai-swarm.md) - Example: a customer service chatbot project
- [Deep Dive into LlamaIndex Workflow: Event-driven LLM architecture](https://www.dataleadsfuture.com/deep-diving-into-llamaindex-workflow-event-driven-llm-architecture.md) - What I think about the progress and shortcomings after practice
- [How to Become a Data Scientist If You Lacking the Necessary Skills](https://www.dataleadsfuture.com/how-to-become-a-data-scientist-if-you-lacking-the-necessary-skills.md) - Embrace Change, Love Learning, and Persist
- [How to Connect LlamaIndex with Private LLM API Deployments](https://www.dataleadsfuture.com/how-to-connect-llamaindex-with-private-llm-api-deployments.md) - When your enterprise doesn't use public models like OpenAI
- [Implementing the train_step Method in Keras 3: From Errors to Solutions](https://www.dataleadsfuture.com/implementing-the-train_step-method-in-keras-3-from-errors-to-solutions.md) - Official code example hasn't kept up with framework updates
- [The Math Behind Keras 3 Optimizers: Deep Understanding and Application](https://www.dataleadsfuture.com/the-math-behind-keras-3-optimizers-deep-understanding-and-application.md) - This is a bit different from what the books say.
- [Keras 3.0 Tutorial: End-to-End Deep Learning Project Guide](https://www.dataleadsfuture.com/keras-3-0-tutorial-end-to-end-deep-learning-project-guide.md) - Implement an encoder-decoder recurrent network from scratch
- [Scikit-learn Visualization Guide: Making Models Speak](https://www.dataleadsfuture.com/scikit-learn-visualization-guide-making-models-speak.md) - Use the Display API to replace complex Matplotlib code
- [Visualizing What Batch Normalization Is and Its Advantages](https://www.dataleadsfuture.com/visualizing-what-batch-normalization-is-and-its-advantages.md) - Optimizing your neural network training with Batch Normalization
- [Ensuring Correct Use of Transformers in Scikit-learn Pipeline](https://www.dataleadsfuture.com/ensuring-correct-use-of-transformers-in-scikit-learn-pipeline.md) - Effective data processing in machine learning projects
- [How to Optimize Multidimensional Numpy Array Operations with Numexpr](https://www.dataleadsfuture.com/how-to-optimize-multidimensional-numpy-array-operations-with-numexpr.md) - A real-world case study of performance optimization in Numpy
- [Exploring Numexpr: A Powerful Engine Behind Pandas](https://www.dataleadsfuture.com/exploring-numexpr-a-powerful-engine-behind-pandas.md) - Enhancing your data analysis performance with Python's Numexpr and Pandas' eval/query functions
- [Seaborn 0.12: An Insightful Guide to the Objects Interface and Declarative Graphics](https://www.dataleadsfuture.com/seaborn-0-12-an-insightful-guide-to-the-objects-interface-and-declarative-graphics.md) - Streamlining your data visualization journey with Python's popular library
- [Efficient k-Nearest Neighbors (k-NN) Solutions with NumPy](https://www.dataleadsfuture.com/efficient-k-nearest-neighbors-k-nn-solutions-with-numpy.md) - Leveraging NumPy’s broadcasting, fancy Indexing, and sorting for performance computing
- [Python Lists Vs. NumPy Arrays: A Deep Dive into Memory Layout and Performance Benefits](https://www.dataleadsfuture.com/python-lists-vs-numpy-arrays-a-deep-dive-into-memory-layout-and-performance-benefits.md) - Exploring allocation differences and efficiency gains
- [Supercharge Your Python Asyncio With Aiomultiprocess: A Comprehensive Guide](https://www.dataleadsfuture.com/supercharge-your-python-asyncio-with-aiomultiprocess-a-comprehensive-guide.md) - Harness the power of asyncio and multiprocessing to turbocharge your applications
- [Conquer Retries in Python Using Tenacity: An End-to-End Tutorial](https://www.dataleadsfuture.com/conquer-retries-in-python-using-tenacity-an-end-to-end-tutorial.md) - Enhancing your Python projects with robust retry mechanisms and error-handling techniques
- [Introducing Python’s Parse: The Ultimate Alternative to Regular Expressions](https://www.dataleadsfuture.com/introducing-pythons-parse-the-ultimate-alternative-to-regular-expressions.md) - Use best practices and real-world examples to demonstrate the powerful text parser library
- [Unleashing the Power of Python Asyncio’s Queue](https://www.dataleadsfuture.com/unleashing-the-power-of-python-asyncios-queue.md) - Mastering the producer-consumer pattern with asyncio through real-life examples
- [Aiomultiprocess: Super Easy Integrate Multiprocessing & Asyncio in Python](https://www.dataleadsfuture.com/aiomultiprocess-super-easy-integrate-multiprocessing-asyncio-in-python.md) - Even no need to know much about asyncio and multiprocessing
- [Mastering Synchronization Primitives in Python Asyncio: A Comprehensive Guide](https://www.dataleadsfuture.com/mastering-synchronization-primitives-in-python-asyncio-a-comprehensive-guide.md) - Best practices for asyncio.Lock, asyncio.Semaphore, asyncio.Event and asyncio.Condition
- [Combining Traditional Thread-Based Code and Asyncio in Python](https://www.dataleadsfuture.com/combining-traditional-thread-based-code-and-asyncio-in-python.md) - A comprehensive guide to integrating synchronous and asynchronous programming in Python
- [Harnessing Multi-Core Power with Asyncio in Python](https://www.dataleadsfuture.com/harnessing-multi-core-power-with-asyncio-in-python.md) - Boost your Python application’s performance by efficiently utilizing multiple CPU cores with asyncio
- [Combining Multiprocessing and Asyncio in Python for Performance Boosts(Updated in 2026)](https://www.dataleadsfuture.com/combining-multiprocessing-and-asyncio-in-python-for-performance-boosts.md) - Using a real-world example to demonstrate a map-reduce program
- [Using Tqdm with Asyncio in Python](https://www.dataleadsfuture.com/using-tqdm-with-asyncio-in-python.md) - An efficient way to monitor concurrent tasks’ progress
- [Use These Methods to Make Your Python Concurrent Tasks Perform Better](https://www.dataleadsfuture.com/use-these-methods-to-make-your-python-concurrent-tasks-perform-better.md) - Best practices for asyncio.gather, asyncio.as_completed, and asyncio.wait
- [Why Taskgroup and Timeout Are so Crucial in Python 3.11 Asyncio](https://www.dataleadsfuture.com/why-taskgroup-and-timeout-are-so-crucial-in-python-3-11-asyncio.md) - Embracing Structured Concurrency in Python 3.11
- [Implement a Cache Decorator with Time to Live Feature in Python](https://www.dataleadsfuture.com/implement-a-cache-decorator-with-time-to-live-feature-in-python.md) - A decorator based on @functools.lru_cache supports cache expiration

## Optional
- [RSS Feed](https://www.dataleadsfuture.com/rss/)
- [Sitemap](https://www.dataleadsfuture.com/sitemap.xml)
- [Full content of pages and posts](https://www.dataleadsfuture.com/llms-full.txt)