Video summary

FlyRank Internship - ML Track - How to start?

Main summary

Key takeaways

Technology

Overview

The speaker explains how to set up and complete ML-track internship assignments using GitHub, Hugging Face, and AI agents (with optional IDE tooling).

Portal / Resources

  • The portal includes an “Announcements” section with frequent updates.
  • There are posts with instructions for using AI models for free on a local machine.
  • Another post focuses specifically on assignments, including weekly tasks.

Weekly Assignment #1: Setup + Running Notebooks

1) Create a GitHub repo from a template

  1. Go to the provided GitHub repository and click “Use this template” / create a new repository.
  2. The repository must be public.
  3. Example name: “FlyRank Internship ML” (spelled similarly to the subtitles).

2) Clone the repo locally

  • Copy the repo URL and clone it inside an IDE.
  • The speaker uses Cursor with Cloud Code / Coders, and suggests OpenCodes/Open Codes may be an easy/cheap option.

3) Create a Hugging Face access token

  1. If needed, create a free Hugging Face account.
  2. Go to: profile → settings → access token → create new token
  3. Enable read permission and name it (example: “internship token”).
  4. Store the token in an environment file (.env) using the variable name exactly: HF token (as mentioned).

4) Verify dataset access

  • The speaker recommends testing with an AI agent to check whether it can access the Hugging Face dataset.
  • They mention you will be auto-approved for requests.
  • Tokens authenticate access to public datasets, but rate limits may apply (not necessarily for simultaneous usage).

5) Run two provided notebooks

  • Notebook locations are available via README links or within the work directory.
  • Run all cells one-by-one to confirm setup and learn the expected workflow/insights.
  • After running, save outputs back to the GitHub repository.

Submission Workflow (GitHub)

  • After finishing a notebook:
    • File → Save
    • Connect/save into the internship GitHub repo created earlier
  • Submission uses the same GitHub URL each time (the repo is already structured with the needed “work folder” contents).
  • Once saved and submitted, it’s considered “done.”

Guidance on Using AI Agents

  • The speaker reassures that using AI agents is not cheating.
  • Still, don’t rely on them solely to finish assignments without understanding.
  • Recommended use:
    • Use agents to explore, understand processes, and interact with datasets
    • Agents can help you learn and may even enable you to “complete assignment 1 or 2” if you already have the necessary ML context (the track teaches this)
  • You can “have conversations” with the dataset/agent, but you should still learn the fundamentals.

Main Sources / Speaker

  • Main speaker: the individual introducing “FlyRank Internship - ML Track - How to start?” and walking through GitHub + Hugging Face + notebook setup.

Original video