Skip to content

QUICKSTART.md

Latest commit

 

History

History
235 lines (179 loc) · 6 KB

QUICKSTART.md

File metadata and controls

235 lines (179 loc) · 6 KB

Quick Start Guide

Get started with your 8-week AI engineering journey in 5 minutes!

Step 1: Review the Plan (5 min)

cd /Users/ankur/D/Playground/ai-engineering-journey

Read these files in order:

  1. README.md - Overview and navigation
  2. PLAN.md - Complete 8-week curriculum
  3. week-01/README.md - This week's detailed plan

Step 2: Set Up Your Environment (10 min)

Create API Keys

You'll need these for Week 1:

  1. OpenAI API Key

  2. Anthropic (Claude) API Key

Environment Setup

# Create .env file in the root
cat > .env << EOF
OPENAI_API_KEY=your-openai-key-here
ANTHROPIC_API_KEY=your-anthropic-key-here
EOF

# Install essential tools (if not already installed)
# Python
brew install python@3.11  # or use pyenv

# Node (for later weeks)
brew install node

# Docker (for databases)
brew install --cask docker

Step 3: Start Week 1 (Today!)

Day 1-2: World View Building

Reading List (6-8 hours over 2 days):

  1. What is Vector Search? - 30 min
  2. OpenAI Embeddings Guide - 45 min
  3. Andrej Karpathy: Intro to LLMs - 1 hour
  4. Browse: Claude Documentation - 1 hour
  5. Explore: ChromaDB basics - 1 hour

Exercise: Create a mental model diagram answering:

  • What are embeddings?
  • How do vector databases work?
  • What is semantic search?
  • How does it all fit together?

Save your diagram in week-01/notes/mental-model.md

Day 3: Start Building

# Navigate to week 1
cd week-01/project

# Create your first project
mkdir semantic-code-search
cd semantic-code-search

# Initialize Python project
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install openai chromadb click rich python-dotenv

# Create basic structure
touch main.py
touch requirements.txt
touch README.md

# Use Claude Code to help you build!

First Task: Build a simple script that:

  1. Takes a text input
  2. Generates an embedding using OpenAI
  3. Stores it in ChromaDB
  4. Retrieves similar texts

Step 4: Daily Workflow

Morning Routine (30 min)

  1. Open week-01/notes/daily-log.md
  2. Review today's goals
  3. Read one article/resource
  4. Plan your coding session

Coding Session (2-4 hours)

  1. Use Claude Code to help you build
  2. But write the core logic yourself
  3. Test as you go
  4. Commit frequently to git

Evening Wrap-Up (15 min)

  1. Update daily log with what you learned
  2. Commit your code
  3. Post on LinkedIn/Twitter (optional but recommended)
  4. Plan tomorrow

Step 5: Leverage Claude Code Effectively

How to Use Claude Code

For Learning:

"Explain how cosine similarity works in semantic search"
"What are the trade-offs between different chunking strategies?"
"Help me understand this error: [paste error]"

For Building:

"Help me scaffold a Click CLI application"
"Review this code for best practices: [paste code]"
"Generate tests for this function: [paste function]"

For Debugging:

"This search isn't returning relevant results. Help me debug."
"How can I optimize this embedding pipeline?"
"What's wrong with my ChromaDB query?"

Golden Rule

  • Let Claude help with scaffolding, debugging, and best practices
  • But understand every line of code
  • Write the core logic yourself to learn deeply

Week 1 Checklist

Setup ✅

  • Environment configured
  • API keys obtained
  • Development tools installed
  • Git repository initialized

Learning ✅

  • Read vector search guide
  • Read embeddings documentation
  • Watched Karpathy video
  • Created mental model

Building ✅

  • Project scaffolded
  • Basic embedding pipeline working
  • ChromaDB integration complete
  • Search functionality implemented
  • CLI interface built

Shipping ✅

  • Code committed to GitHub
  • Documentation written
  • Blog post drafted
  • Shared on social media

Tips for Success

Time Management

  • Block time: Dedicate specific hours each day
  • Eliminate distractions: Close Slack, email, etc.
  • Take breaks: Pomodoro technique works well
  • Don't perfectionism: Ship and iterate

Learning Approach

  • Build to learn: Don't just read, code
  • Explain to understand: Write blog posts
  • Ask questions: Use Claude Code actively
  • Reflect daily: What did you learn?

Staying Motivated

  • Share progress: Post daily updates
  • Celebrate wins: Shipped a feature? Tweet it!
  • Join communities: Discord servers, Reddit
  • Visualize the goal: Your portfolio in 8 weeks

Common Pitfalls to Avoid

❌ Tutorials hell: Don't just watch, build ❌ Perfectionism: Ship MVP, iterate later ❌ Scope creep: Stick to the week's project ❌ Isolation: Share your journey, ask for help ❌ Burnout: Take breaks, pace yourself

✅ Do this instead: Build → Ship → Learn → Repeat

Getting Help

When Stuck

  1. Google it - You're not the first!
  2. Ask Claude Code - Your AI pair programmer
  3. Check documentation - Primary source of truth
  4. Stack Overflow - Community wisdom
  5. Discord/Reddit - Active communities

Resources

  • See resources/README.md for comprehensive links
  • Check Week 1 README for week-specific resources
  • Use PROGRESS.md to track your journey

Ready to Start?

  1. Right now: Read Pinecone's vector search guide
  2. Today: Complete Day 1 reading
  3. This week: Ship your first AI tool
  4. Next 8 weeks: Transform into an AI engineer

Remember: You're a Director of Engineering with strong systems thinking. You've got this! The goal isn't perfection—it's progress, learning, and shipping.

Let's build! 🚀

Start with: week-01/README.md