Browse 20 learning paths
AI / Intelligent Development
AI Development Directory
A fast-reference directory of AI agents, coding assistants, and machine learning frameworks. Each entry is a short, practical primer -- what the tool is, what it is actually good for, and the most common mistake people make evaluating it -- so you can get oriented before going deep on any single one.
11 tutorials
Artificial Intelligence: A Practical Starting Point
Place machine learning, deep learning, and agents inside one field, learn why every deployed system is narrow, and route yourself to the right tool for the problem in front of you.
Natural Language Processing: Where AI Meets Language
Sort text problems into the task families that have known solutions, and choose between a small specialised model and a general language model on cost, latency, and control.
Scikit-learn: Classical Machine Learning in Python
Use one consistent interface across every classical algorithm, put every preprocessing step inside a pipeline so validation stays honest, and know why tabular problems still start here.
PyTorch: A Flexible Deep Learning Framework
Read a training step line by line, understand what autograd is tracking, and know when to move from eager execution to a compiled or exported graph.
TensorFlow: Google's Machine Learning Framework
Understand what graph execution buys you, use the Keras layers that most work happens at, and pick the deployment target — server, browser, phone, or microcontroller — that actually drives the choice.
OpenCV: Computer Vision Building Blocks
Use classical vision for the parts of a pipeline that do not need learning, avoid the colour-space and coordinate bugs that catch everyone, and know where a trained model has to take over.
LangChain: A Framework for LLM Applications
Know which pieces of an LLM application are genuinely worth a framework, what the retrieval path actually does, and when the abstraction costs more than the plumbing it replaces.
Ollama: Running Open LLMs Locally
Size a local model against the memory you actually have, read quantisation labels correctly, and know which workloads belong on your own hardware rather than a hosted API.
Hermes Agent: An Open, Self-Improving AI Agent
Read the self-improvement claim precisely, understand what a learned skill actually is, and scope an open self-hosted agent's permissions before it touches anything real.
Selenium: Automating the Browser
Understand what WebDriver actually controls, fix flakiness at its four real sources, and write selectors that survive a redesign.
Playwright: Modern End-to-End Testing
Understand exactly what auto-waiting checks before a click, use network interception to make tests deterministic, and read a trace instead of guessing why CI failed.