🐍 Section 1: Python 101
5 weeks of Python fundamentals. Choose a Normal track (no installs, code right in your browser) or a Hard track where you build a tiny language model from scratch using nothing but built-in Python.
Explore Python 101📊 Section 2: Pandas & Data Analysis
5 weeks of pandas and exploratory data analysis. Reproduce a Kaggle-style notebook on the Normal track, or run a full EDA project with visualizations on the Hard track.
Explore Data Analysis🌍 Real-World Projects
Practical projects you can build with Python, installed for real on your own machine — browse any time, no need to finish the course first.
Build a Dependency-Freshness Checker
August 2027
A real CLI that reads a pyproject.toml, checks PyPI for newer dependency versions, and reports what's outdated — no API key needed.
Build a Git Commit-Message Generator
July 2026
Draft a Conventional-Commits-style message from a real staged git diff with a free-tier LLM, and only commit it after you explicitly confirm.
Query a Database in Plain English with MCP
August 2027
Build an MCP server that exposes a local SQLite database, then watch an LLM client write and run its own SQL to answer plain-English questions about it.
Build a Discord Trivia Bot
July 2026
Run trivia rounds in a Discord server with discord.py: a persistent leaderboard, and questions generated fresh on any topic with a free-tier LLM.
Build an MCP Server for Your Notes
July 2026
Index a real folder of Markdown notes and expose it to Claude Desktop as searchable tools with the Model Context Protocol.
Build a Recipe-Planner Agent
July 2026
Build a tool-using AI agent that suggests meals from ingredients you have on hand, grounded in a real local recipe database instead of guessing.
Build a Meeting-Notes Summarizer
July 2026
Turn a raw meeting transcript into structured decisions, action items, and open questions using a free-tier LLM and a carefully designed JSON-extraction prompt.
Build a GitHub Issue Triage Agent
July 2026
Fetch open issues from a real public GitHub repo and use a free-tier LLM to draft suggested triage labels for a human maintainer to review.
Build a Voice-to-Task Agent
July 2026
Transcribe a voice memo locally and for free with OpenAI's open-source Whisper model, then use a free-tier LLM to turn it into a structured task list.
Build a Study-Buddy Quiz Agent
July 2026
Turn your own study notes into an interactive quiz: a free-tier LLM writes questions grounded in your notes, then judges your typed answers.
Turn a Codebase into a Knowledge Graph
July 2026
Parse a real codebase's Python files with the ast module, build a graph of its structure with networkx, and visualize and query it — no API key needed.
Build a RAG-Backed Docs Q&A Discord Bot
July 2026
Wrap the RAG App project's retrieval pipeline in a live Discord bot that answers questions from a folder of documentation.
Build a Personal Email-Triage Agent
July 2026
Categorize, prioritize, and draft replies for a batch of emails with a free-tier LLM — the agent never sends anything, only you do.
Build a Multi-Agent Research Assistant
July 2026
Build a small multi-agent system — a planner, a researcher, and a writer — that breaks down a research question and synthesizes a final report, using deepagents sub-agents and a free-tier LLM.
Build an Agentic Code Reviewer
July 2026
Build a CLI tool that reads a real git diff via subprocess and asks a free-tier LLM to review it like a human reviewer would.
Build an MCP Server
July 2026
Build a Model Context Protocol server exposing your own tools, and connect it to a real AI client like Claude Desktop.
Train Your First Machine Learning Model
July 2026
Go from describing data to predicting from it: train and compare binary classifiers with scikit-learn on the Titanic dataset.
Build a RAG App Over Your Own Notes
July 2026
Chat with your own notes: local embeddings with sentence-transformers, NumPy similarity search, and a free-tier LLM for the final answer.
Chat with Your PDFs
August 2027
Multi-document RAG over a folder of PDFs: page-aware chunking, local embeddings, and page-cited answers from a free-tier LLM.
Scrape and Analyze a Live Website
July 2026
Scrape a real, scraping-friendly website with requests and BeautifulSoup, then clean and chart the results with pandas and matplotlib — no API key needed.
Build a Job-Listing Aggregator
August 2027
Scrape multiple job-board-style sources, dedupe listings across them, and alert on new matches against a keyword filter — no API key needed.
Fine-tune a Small Language Model
July 2026
Fine-tune a small open-source language model with LoRA using Unsloth, on a free Colab/Kaggle GPU.
Build an AI Agent
July 2026
Install Python for real and build your first AI agent with LangChain's deepagents, using a free-tier API key.
Build a Rate-Limited API Service
August 2027
Build a real FastAPI service wrapping your own dataset, with API-key auth and a sliding-window rate limiter built from scratch.
Build a Personal Finance Agent
July 2026
Categorize a bank CSV export and flag spending anomalies, combining pandas with an LLM tool-calling agent.
Build a Browser-Automation Agent
August 2027
Combine Playwright browser automation with a free-tier LLM tool-calling agent that fills out a real practice web form on its own.
Count Objects in Real Time with a Webcam
July 2026
Count objects live from a webcam feed with OpenCV and a pretrained YOLO11n model, no API key or signup needed.
Build a Wordle Clone
August 2027
Build a real terminal Wordle game from scratch: correct guess feedback, a custom word list, and persistent stats tracking across sessions.
Build a Habit-Streak Visualizer
August 2027
Track daily habit check-ins locally and render a GitHub-contributions-graph-style calendar heatmap, for any habit you want to track.