Home / Start Here

// Path 06 — 5-level beginner curriculum

New to AI?
Start here.

You do not need to know everything before you start. This path is for professionals with no technical background in AI who want to build real skills, not just vocabulary. Pick it up and build something this week.

Start with day 1 → All paths

// The core principle

"You do not need to know everything before you start."

Most professionals who look like AI experts learned most of it in the last 18 months. The window is open.

— virat-lab / beginner-path

// Week 1 — do this now

7 days.
7 concrete tasks.

Check each task off as you complete it. Your progress saves automatically. Come back tomorrow and pick up where you left off.

0 / 7 complete

Day 1

Pick your path

Browse the 6 learning paths. Choose the one closest to your current role or where you want to go. Write one sentence about why you chose it.

Action: choose a path at viratlabs.com/#paths

Day 2

Learn the vocabulary

LLM. Prompt. Context window. Hallucination. Token. Tool use. Look each one up. Write a one-sentence definition in your own words.

Action: write 6 definitions. No looking back once written.

Day 3

Write 5 prompts

Use the Role-Goal-Context-Format template. Apply it to 5 real tasks from your current job. Test each one in Claude or ChatGPT.

Action: save your best prompt. That is your first artifact.

Day 4

Map a workflow

Pick one repeatable task you do every week. Map it step by step. Mark each step [manual] or [AI could assist]. Estimate time saved.

Action: one workflow map. You now have a design document.

Day 5

Learn 5 risks

Prompt injection. Hallucination. Data leakage. Overconfidence. Privacy violation. For each: write what it is and one example from your field.

Action: one page. These are the risks you own.

Day 6

Research with AI

Find a paper, article, or report in your field. Use AI to summarize it. Then verify the summary yourself. Document any errors the AI made.

Action: build the habit of verify-after-use. Always.

Day 7

Build your first artifact

Write one Markdown document — your best prompt, your workflow map, or your risk one-pager. Publish it somewhere. Even a private note counts.

Action: something exists now that did not before. Ship it.

// The full 5-level curriculum

Five levels.
90 days. One artifact each.

Each level has a concept to understand and something to build. The builds are your portfolio — real evidence of capability, not credentials.

Level 1

Core Terminology

An accurate mental model before anything else. Most people skip this and pay for it later.

↳ What to learn

AI: software that performs tasks normally requiring human intelligence
LLM (Large Language Model): what it is, how it generates output, why it is probabilistic
Prompt: the instruction you give an AI model
Context window: how much the model can "see" at once, and why it matters
Hallucination: when the model produces confident, plausible, false output
Tool use: when AI calls external tools (search, APIs, files) to complete a task

↳ What to build

Vocabulary document — 6+ AI terms defined in your own words, with one example from your field for each. No looking at your notes when you write. This forces actual understanding, not memorization.

Level 2

Practical Prompting

The Role-Goal-Context-Format template applied to real tasks. This is the skill that pays immediately.

↳ What to learn

Role: setting the AI's expertise and decision-making lens
Goal: one clear, specific objective per prompt
Context: what the AI needs to know — most under-specified part
Format: exactly how you want the output structured
Common failure modes: vague goals, no format, missing context
How to iterate: what to change when the output is not what you wanted

↳ What to build

Prompt library — 10 prompts — One prompt for each of 10 real tasks from your job. Each prompt should be tested and annotated with: what worked, what did not, and how you improved it.

Level 3

Workflows

A workflow is a repeatable process. This level teaches you to design, document, and run AI-assisted workflows.

↳ What to learn

What makes a workflow: trigger, inputs, steps, output, human review
How to identify which steps benefit from AI assistance vs. human judgment
The 6-step workflow improvement cycle: collect → prompt → review → improve → document → repeat
How to document a workflow so someone else could run it

↳ What to build

One documented workflow — A repeatable AI-assisted workflow from your actual work. Include: trigger, inputs, steps with which are AI-assisted, output, who reviews it, and estimated time saved.

Level 4

Safety & Critical Evaluation

Knowing when not to trust AI is the skill that separates good practitioners from careless ones.

↳ Safety rules to internalize

Protect sensitive information — do not share what you would not share publicly
Verify claims — especially statistics, citations, dates, and names
Encourage uncertainty — ask the model to flag what it is not sure about
Get human review for high-stakes domains — medical, legal, financial decisions
Treat AI output as a draft — not a final product

↳ What to build

Risk checklist for your role — A one-page checklist specific to your job: what data is safe to use with AI, what decisions always require human review, and what to do when AI gives you a confident wrong answer.

Level 5

Portfolio Project

One tangible artifact that demonstrates what you have learned. This is the output that matters.

↳ What your project must document

What problem you solved or process you improved
Who benefits from this (your team, your users, yourself)
How it works — the prompts, the workflow, the tools
What risks you identified and how you mitigated them
Where human review is required in the process

↳ Portfolio project options

AI-assisted workflow — A fully documented, repeatable workflow from your job that uses AI at one or more steps, with safety controls and a human review gate
Prompt library for your team — 15+ tested, annotated prompts specific to your team's work, with a short guide on how to use and improve them
AI risk assessment — A thorough risk assessment of AI tool use in your organization or department, with recommendations

// Your first prompt

Copy this.
Use it today.

This is the starter prompt from the beginner path. It uses the Role-Goal-Context-Format template from Level 2. Fill in the brackets and paste into Claude, ChatGPT, or any AI tool.

After you use it

Ask yourself: was the output useful? What would make the prompt better? Change one thing and try again. That iteration is the practice.

// Role-Goal-Context-Format starter

Role: You are a thoughtful AI learning advisor.

Goal: Help me figure out which AI skill I should
learn first, given my current job and goals.

Context: My current role is [your job title].
My main daily tasks are [list 3-5 things you do
at work every week].
I have [no / some / strong] technical background.
My goal is to [use AI at work / change careers /
understand AI better].

Format: Recommend one specific AI skill to focus
on first. Explain in 2-3 sentences why it fits
my situation. Give me one concrete thing I can
try this week using a free AI tool.

// After level 5

Choose your
specialization.

After building the foundation, you will have a much clearer sense of which direction fits your role and goals. These five paths are where the real career opportunities are.

🔐

Cybersecurity + AI

3-level curriculum. For security professionals.

🏥

Healthcare + AI

5-phase curriculum. For clinical professionals.

⚡

Automation + AI

8 use cases. For ops and workflow builders.

🤖

AI Agents

5-phase roadmap. For developers and engineers.