Video summary

Free AI Engineer Course | 8 Weeks | Episode 01: Intro + Setup

Main summary

Key takeaways

Educational

Main Ideas / Lessons

Course purpose (2026 AI Engineer role)

  • The speaker, Pratyush Narayan, introduces a “Day One” episode designed to help viewers land an AI Engineer / Agentic Engineer role in 2026.
  • The course exists because most AI courses fall into two unhelpful extremes:
    1. Math-heavy / research-heavy (e.g., transformer internals), which help for deep research/PhD paths but not for interview-ready skills.
    2. Tutorial-style “wrapper” courses that feel productive but teach little, leaving learners unable to answer interview questions meaningfully.
  • This course is designed to be interview-relevant, built by reverse engineering the questions asked in AI fresher/experience interviews.

What the course will teach (high-level outcomes)

  • Duration: 8 weeks (about 40–45 episodes, each 30–60 minutes)
  • By the end, learners can cover interview topics and build a production-style end-to-end AI project, including:
    • LLM fundamentals
    • RAG and Document Q&A
    • Agents and autonomous research agents
    • LangGraph
    • MCP server
    • Observability, evaluation, and deployment
    • Fine-tuning
  • Emphasis: “what to know for interviews,” not PhD-level ML/AI theory.

Who the course is for / not for

For:

  • College students with little/no AI background
  • Working professionals with ~1–3 years experience who want to switch into AI
  • People in service companies transitioning into AI development/projects

Not for:

  • Anyone specifically targeting a PhD
  • Anyone wanting deep math / deep learning logic
  • Those looking to study Machine Learning academically (the speaker explicitly says they won’t teach ML that way)

Prerequisites (learning requirements)

  • Basic Python knowledge is required.
    • Option A: Use the speaker’s playlist “DSA Basics in Four Days” and watch the Python part.
    • Option B: Learn only enough Python (e.g., variables, dictionaries, lists/tuples). No mastery required.

Methodology / Instruction Steps (Episode 01: Setup Plan)

A) Learning structure

  • Episode 01 (Day One) is setup only:
    • Open a laptop and follow guided setup steps.
    • Real content starts in Episode 2.
  • Setup is intended to be manageable and guided via documentation/screen recordings.

B) Model/runtime approach choice (local vs. cloud)

  • The course avoids relying on Ollama for everyone because:
    • Ollama requires downloading and running a local LLM.
    • It can be difficult on low-RAM laptops (e.g., ~4GB).
  • Instead, it uses:
    • Cloud-based Grok (referred to as “Gok/Grok” in subtitles)
    • Quadrant (later for vector DB / RAG and related tasks)
  • Rationale: cloud setup prevents local model issues and makes learning smoother.

C) Setup requirements

Minimum:

  • A laptop
  • Internet
  • A rough notebook (to write secret keys)
  • Earlier steps include account creation (the overall message is that the course is free, despite an ambiguous “₹1” mention)

D) Install basic tools (Windows-first)

1. Install Python

  • Use the provided Python installer link from the description.
  • Choose the correct OS download (Windows/Mac options are mentioned).
  • Run the installer:
    • Click Install Python
    • If installation fails, stop and rerun
    • When prompted, keep “Add to PATH” checked
  • Verify:
    • Open PowerShell (Windows + R → PowerShell)
    • Run: python --version

2. Install Visual Studio Code (VS Code)

  • Download using the link in the description.
  • Accept the agreement and use defaults.
  • Ensure “Add to Path” is not unchecked during installation (re-emphasized due to Windows path issues).
  • Verify you can open VS Code.

3. Install VS Code extensions (Microsoft)

Install:

  • Python extension
  • Pylance
  • Jupyter

(Subtitles mention “three extensions: Python, PylS/PyS, and Jupyter,” with context indicating Microsoft-published Python/Jupyter-related tooling.)

4. Install Git (not GitHub)

  • Download Git using the provided link.
  • Verify installation:
    • git --version (in PowerShell)
  • Configure Git identity next.

5. Install UV

  • Run the provided UV command to install it.
  • UV is described as a faster package installer, similar to pip.
  • Reopen terminal/PowerShell and verify:
    • uv --version

6. Install Microsoft Terminal (optional but recommended)

  • Install from Microsoft Store.
  • Search for: “Microsoft Terminal”

E) Account setup and API key creation (secrets handling)

1. If you don’t have a GitHub account

  • Create a GitHub account.
  • Note your username/email.

2. Configure Git username + email

Run:

  • git config --global user.name ...
  • git config --global user.email ...

Verify:

  • git config --global -l

3. Create/obtain Grok API key (cloud Grok)

  • Log into the Grok cloud site using Google.
  • In API keys:
    • Create an API key with a chosen name
    • Copy and store it immediately (it’s only visible once)
  • Strong emphasis: save it in your notebook/notes and don’t close until the recording is done.

4. Create/obtain Quadrant access

  • Log into Quadrant using Google.
  • Create a free cluster (no paid spend required).
  • Copy and save:
    • Quadrant API key
    • Cluster endpoint URL
  • Strong emphasis: both are treated as secrets and must be saved.

F) Save API keys into an environment file (Windows)

  • Create a local project folder and a .env file to store secrets.

Steps (as described):

  1. Open PowerShell.
  2. Go to your user directory:
    • cd C:\Users\<your-username>\
  3. Create a project folder (example):
    • mkdir <your-folder-name>
    • cd <your-folder-name>
  4. Create a .env file:
    • Use Notepad: notepad .env
    • (If it doesn’t exist, Notepad creates it.)
  5. Paste values into the .env file as assignments:
    • Grok API key line:
      • GROK_API_KEY = <your-grok-key>
    • Quadrant endpoint:
      • QUADRANT_URL = <your-quadrant-cluster-endpoint>
    • Quadrant API key:
      • QUADRANT_API_KEY = <your-quadrant-api-key>
  6. Repeated stress from the speaker:
    • No spaces
    • No commas
    • No extra quoting/symbols
  7. Save with Ctrl + S.

Finally:

  • Open the folder in VS Code.
  • The folder will include:
    • your Week folder(s)
    • the .env secret file
  • Treat .env as secret and don’t push it to Git.

G) Final verification (“checks” before continuing)

Confirm installations/versions are correct by running:

  • python --version
  • uv --version
  • git --version
  • Ensure Git is configured with a usable Git identity
  • (Also mentions confirming VS Code, possibly via code --version)

If all checks pass:

  • Episode 01 is complete.

Speakers / Sources Featured

Speaker

  • Pratyush Narayan
    • IIT graduate
    • Works at Akamai
    • Introduced as the instructor

Referenced services/tools (used in setup)

  • Grok (cloud Grok)
  • Quadrant (QuadrantIO)
  • Git / GitHub (account and Git identity)
  • Python
  • Visual Studio Code (VS Code)
  • VS Code extensions (Python / Pylance / Jupyter—Microsoft-published)
  • UV (package installer)
  • Microsoft Terminal (optional)
  • ChatGPT (mentioned as a troubleshooting assistant for installation issues)

Original video