Day 1: Starting the "Hello Agents" Open Source Guide

November 23, 2025 - Day 1: Starting the "Hello Agents" Open Source Guide

Today I explored an excellent open-source resource on GitHub called Hello Agents, created by the Datawhale community. There are many materials about AI agents online, but this one stands out — it's clear, well-structured, and highly practical.

🔗 GitHub link: https://github.com/datawhalechina/hello-agents (opens in a new tab)

The project contains 16 chapters organized into 5 main sections. Each section includes code samples, explanations, and hands-on exercises — exactly the kind of structured practice I've been looking for.


🚀 Study Goal

I set a goal to finish the full guide in 5 days, meaning roughly one main section per day.

Today, I completed Section 1, which covered the history of agents, LLM basics, core NLP concepts, encoder-decoder architectures, prompt engineering, and tokenization fundamentals.

Key insights gained:

  • Understanding the evolution from rule-based systems to modern LLM-powered agents
  • Learning about transformer architecture and how it enables better long-range dependencies compared to RNNs
  • Comparing different NLP approaches (n-grams, RNNs, transformers) and their use cases
  • Understanding encoder-decoder patterns and attention mechanisms
  • Learning effective prompt engineering techniques and best practices
  • Understanding tokenization strategies and model size trade-offs

Action items completed:

  • Reviewed transformer attention mechanisms
  • Practiced with different model sizes
  • Understood tokenization vs. character-level approaches
  • Learned prompt engineering best practices
  • Compared NLP approaches for different use cases

This systematic approach through the Hello Agents guide provided an excellent foundation for understanding modern AI agent systems.


Overall Reflection: The Hello Agents material provided an exceptionally clean and systematic summary of Section 1 topics. This structured approach helped reinforce the mental framework behind LLMs and agents, making complex concepts more digestible and actionable.


🎯 My Learning Intention

🎯 Mood📊 Progress💡 Key Takeaway🎯 Tomorrow's Goal
Excited and focused1/5 sections completed (20%)Structured open-source learning resources are invaluable for building technical depthComplete Section 2 of Hello Agents

Progress Bar: ■■■□□□□□□□ (20%)

🎯 My Learning Intention

I'll continue studying daily and review the included interview questions to deepen my understanding. My goal is to build an intuitive, explainable grasp of agent and LLM fundamentals — the kind of understanding that allows me to break concepts down for anyone.

This is an important step toward becoming a strong AI product manager with solid technical depth.

On to Day 2 tomorrow!