Free 2-hour course by Andrew Ng and Isa Fulford. Covers zero-shot, few-shot, chain-of-thought prompting, iterative refinement, and building applications with the OpenAI API.
Install Cursor IDE and connect it to your Claude or OpenAI API key. Work through the DeepLearning.AI course and write 10 prompts that solve real coding tasks you face at work. For each prompt, note: what worked, what failed, and why.
Robert C. Martin's foundational rules for writing readable, intention-revealing code. These three chapters alone will permanently change how you name variables, size functions, and write comments.
Pick one real module or file from a codebase you work on. Apply the naming and function-size rules from Clean Code Ch. 1–3. Commit the before and after as two separate commits. Write a short comment in the commit message explaining each change.
Practical guide to technical writing by Jared Bhatti et al. The first two chapters cover why documentation matters, who your audience is, and what makes a README actually useful.
Write or completely rewrite the README for one of your projects. Include: what the project does, how to set it up, how to use it, and one concrete usage example. Share it with at least one peer and ask: "Is anything unclear?"
Free 2.5-hour course by Isa Fulford and Andrew Ng. Covers chaining prompts, building multi-step pipelines, evaluation frameworks, and best practices for production LLM systems.
Build a CLI tool that uses an LLM API to automate a dev task you do regularly. Good options: auto-generate commit messages from git diff, summarize PR descriptions, or answer questions about a codebase file. Commit to GitHub with a working README.
David Thomas and Andrew Hunt's foundational guide for software craftspeople. Chapter 1 covers responsibility and avoiding broken windows. Chapter 2 covers DRY, orthogonality, tracer bullets, and prototyping.
Scan a codebase you own or contribute to and identify 3 "broken windows": messy code, dead code, misleading names, or known tech debt that nobody has fixed. Fix each one. Document what was wrong, what you changed, and which Pragmatic Programmer principle applies.
GitLab's publicly available guide on async-first communication, written decisions, and reducing meeting dependency. Essential reading for distributed and remote developer teams.
Record a Loom (or equivalent) video of at most 3 minutes showing your AI CLI tool to a non-technical colleague. Focus on the problem it solves, not the code. Share the link with at least one person and ask for one line of feedback.
Free 3-hour course by Harrison Chase and Andrew Ng. Covers LangChain models, prompts, parsers, memory, chains, and question-answering over documents — the building blocks of real LLM apps.
Extend the CLI tool from Week 2 using LangChain. Add one of: a second chained prompt step, conversation memory across turns, or a document Q&A feature over a local file. Push to GitHub and update the README with what changed and why.
Robert C. Martin's definitive talk on the five SOLID principles: Single Responsibility, Open/Closed, Liskov Substitution, Interface Segregation, and Dependency Inversion. Free on YouTube, ~1.5h.
Draw a simple architecture diagram for one of your existing projects (boxes and arrows is fine, use Excalidraw or draw.io). Label the components. Then identify at least 2 places that violate one of the SOLID principles and write a short note on how you would fix each.
Austin Kleon's practical guide on sharing your work publicly as a creative professional. Chapters 1–4 cover the mindset shift from hiding to sharing, the power of process documentation, and building an audience by being useful.
Write and publish a LinkedIn post of at least 200 words about something you built or learned in this training so far. Include a screenshot, diagram, or code snippet. Do not wait for it to be perfect — publish and move on.
Free 2-hour course covering OpenAI function calling, LangChain tools, routing, and building agents with ReAct. The natural next step after LangChain for LLM Application Development.
Do a final pass on your AI tool: add proper error handling for the main failure cases, write a test for the core happy path, and clean up the README. Then deploy it (even a simple GitHub release) or share it in a dev community — Slack, Discord, or a GitHub discussion.
Martin Kleppmann's definitive book on modern system design. Chapter 1 defines the three pillars every senior developer is expected to reason about: reliability, scalability, and maintainability.
Write a 1-page design doc for a system you would build to solve a real problem you have encountered. Include: problem statement, key constraints, main components and their responsibilities, and at least one trade-off you consciously made. Use Kleppmann Ch. 1 vocabulary: reliability, scalability, maintainability.
John Sonmez's guide to treating a developer career as a business. The first five chapters cover career positioning, finding your niche, personal branding, and setting professional goals.
Using AI assistance, rewrite your GitHub profile README and your LinkedIn About section. They should clearly state what you build, what you are learning, and what you are looking for. Publish one public 3-month learning goal — in a LinkedIn post, a pinned GitHub gist, or your profile README.