How to use this roadmap: Expert Workflows is the project-based track that follows AI Essentials and AI Deep Dive. This page lists only complete, currently published lessons. Each lesson should leave you with a working local prototype and a clear verification step.
What this track covers
The current track contains two finished projects: a Python PDF summarizer and a Chrome extension. Start with the PDF project if you want to practise files, API calls, chunking, and structured output. Start with the extension project if you already understand API basics and want to package a browser tool.
Examples are reference implementations. Check provider documentation for current model access, pricing, quotas, permissions, and data-handling rules before using a project with private or organisational information.
Project 1: Build a Python PDF summarizer
This four-part project progresses from environment setup to a reusable summarization workflow.
- Set up the environment and make the first API call — configure Python, keep credentials outside source code, and confirm the connection.
- Extract text from PDFs and identify scan limitations — handle text-based files and recognise when OCR is required.
- Chunk long documents and handle token limits — split input, preserve context, and reduce partial summaries.
- Validate structured output and package the tool — check JSON before use and prepare a safer reusable project.
Project 2: Build a practical Chrome extension
This three-part project starts with a minimal extension, then adds page interaction and a usable task interface.
- Create an extension that loads correctly — build the manifest and popup, then verify the extension locally.
- Read page content and add an AI-assisted action — understand permissions, page access, privacy boundaries, and credential risks.
- Add task buttons, icons, and a repeatable workflow — refine the interface without expanding permissions unnecessarily.
Before sharing or deploying a project
- Remove API keys, private documents, and local configuration from shared files.
- Use a server-side credential boundary for any tool distributed to other people.
- Review extension permissions and disclose what page data is read or transmitted.
- Test with your own non-sensitive sample files and record the environment and date.
- Verify AI output against the source instead of treating a summary as a replacement for the document.
Choose your next step
New to APIs and document processing? Begin with the PDF summarizer setup. Comfortable with those basics? Begin with the Chrome extension project.
Last reviewed: August 2026. Written and edited by G. Troy. AI tools may assist drafting and organisation; technical claims and links are reviewed before publication.