I’m in high school learning Python with AI help. What should I learn next, when should I use GitHub, and is a caption-generator video project a good portfolio idea?

Short Answer

In short, focus on these checks: roadmap, video project idea, and using AI wisely. Avoid rebuilding the whole content strategy from one trend, tool, or metric; run a small test and compare the audience response before making more videos. The caveats are platform context, production effort, content quality, and whether the next step is measurable. More detailed guidance is below.

Yes, a caption-generator or short-video helper can be a good portfolio project, but only after you can build and explain a small web app without depending on AI for every step. Your current projects are enough to prove you are practicing. The next goal is not “more tutorials”; it is learning how to finish, document, test, and publish one useful project end to end.

For the next few months, stay boring on purpose: Python fundamentals, Git and GitHub, one web framework, basic databases, HTML forms/templates, APIs, and deployment. Then build one portfolio project that solves a real workflow, such as turning a transcript into timed captions for a vertical video, reviewing the captions, and exporting a simple caption file.

Roadmap

First, finish enough Flask to understand routes, templates, forms, sessions, database models, and app structure. You do not need to memorize every decorator or copy every CSS line perfectly. You do need to understand what the browser sends, what the server returns, where data is stored, and how errors are handled.

Second, start GitHub now, not later. GitHub’s own beginner guide teaches repositories, branches, commits, and pull requests because those are normal development habits, not “big project” habits. Put even small projects there, but clean them first: add a README, screenshots or a demo video, setup steps, what you learned, and what you would improve.

Third, pick one serious project per learning block. For example: after Flask, build a habit tracker with accounts and charts; after databases, build an expense dashboard; after APIs, build a YouTube-style caption helper; after machine learning basics, build the hand-sign detector. The dream projects are allowed early, but reduce them to version one. A voice-controlled chess project can start as “recognize three commands and move pieces in a local board,” not “full speech AI chess assistant.”

Video Project Idea

A strong beginner-to-intermediate portfolio project would be a simple web app where a user uploads a short clip or pastes a script, creates timed captions, reviews them, and exports WebVTT or SRT. If you want a faster prototype, draft a captioned demo in YumCut first so you can study the user flow before coding your own small version.

This project is useful because it touches several real skills: file handling, text processing, timestamps, frontend previews, accessibility, and product thinking. W3C guidance is clear that automatic captions are only a starting point and usually need human correction, so your app should include a review screen instead of pretending AI output is final.

Keep the first version simple. Do not start with perfect speech recognition, translation, fancy animations, or accounts. Start with a text transcript, let the user split it into caption lines, assign rough timestamps, preview the result, and export a caption file. Then add AI assistance later: suggest line breaks, shorten captions, translate into one extra language, or generate a short social caption from the video topic.

Using AI Wisely

AI is fine as a tutor, but it should not become your steering wheel. Ask it to explain errors, quiz you, review your code, and compare two approaches. Do not let it choose your entire roadmap every week. A good test is simple: can you close the AI chat and rebuild the smallest version from memory? If not, slow down and write notes.

For content or portfolio demos, Google Search Central’s people-first guidance is a useful mindset: make something because it helps a real user, not because it sounds trendy. If you use AI heavily, be ready to explain what it did, what you checked manually, and why the result is still useful. You can test the workflow in YumCut, but your portfolio write-up should show your own engineering decisions.

Manual review matters most where the output affects meaning: captions, translations, accessibility, educational explanations, finance or health claims, and anything shown publicly. AI can draft quickly; you still need to verify timing, spelling, names, numbers, and whether the final clip makes sense without extra context.

What Employers Notice

For a student, employers and mentors do not expect huge production systems. They notice clean fundamentals: readable code, small functions, meaningful commits, simple tests, clear README files, and projects that actually run. A finished small project is more convincing than ten half-finished ambitious ones.

Your communication is also part of the skill. When asking for help, lead with your goal, current level, what you tried, where you got stuck, and the exact question. That makes people more likely to answer and trains the same clarity you will need in college projects, internships, and job interviews.