Learn Lab 🧪
Welcome to the PyRun Learn Lab — a hands-on playground where you move from "I have no idea how the cloud works" to "I just ran 100 serverless functions in parallel to analyze a genome."
This section is designed for two types of explorers:
- Students & Researchers who have never touched AWS and want to learn cloud concepts by doing, guided by AI.
- Data Scientists who know their pandas, NumPy, and scikit-learn pipelines by heart, but have never distributed them across thousands of cores in the cloud.
How It Works
Each experiment follows the same narrative arc:
We show you everything:
- The original prompt we sent to the AI
- The local code that only runs on your laptop
- The generated cloud code that scales infinitely
- Screenshots of the actual execution
- Mermaid diagrams explaining the architecture
Experiment Catalog
🎓 Student Cloud Basics
Perfect if you have never used AWS. These are the "Hello World" moments of the cloud.
| Experiment | What You Learn | Difficulty |
|---|---|---|
| S3 File Writer | Writing files to Amazon S3 from a Lambda function | ⭐ |
| DynamoDB Counter | Building a serverless API with API Gateway + DynamoDB | ⭐⭐ |
🗳️ Pet Vote App
A complete full-stack serverless application. You migrate a local Flask app (HTML + Python) to AWS: S3 static hosting, API Gateway, Lambda, and DynamoDB.
| Experiment | What You Learn | Difficulty |
|---|---|---|
| Pet Vote | Full-stack migration: Frontend → API → Database | ⭐⭐⭐ |
🧬 DNA Analysis (Distributed Genomics)
For data scientists. You take a local bioinformatics script that crashes on large FASTQ files and distribute it across 100 AWS Lambda functions using Lithops + DataPlug.
| Experiment | What You Learn | Difficulty |
|---|---|---|
| DNA Analysis | Zero-data-movement partitioning + MapReduce serverless | ⭐⭐⭐⭐ |
🛰️ DANA Flood Analysis (Distributed Geospatial)
For geospatial researchers. You analyze satellite imagery of the Valencia DANA floods at 30m resolution by distributing spectral index computation across AWS Lambda.
| Experiment | What You Learn | Difficulty |
|---|---|---|
| DANA Flood Analysis | Hybrid local/cloud geospatial pipelines at scale | ⭐⭐⭐⭐ |
📈 Serverless Stock Pipeline (Event-Driven Architecture)
For students ready to connect multiple services. You build a real data pipeline where a local producer feeds a cloud message queue, and a Lambda function automatically stores data in DynamoDB.
| Experiment | What You Learn | Difficulty |
|---|---|---|
| Stock Pipeline | Local producer → SQS → Lambda → DynamoDB event-driven flow | ⭐⭐⭐ |
🌡️ Global Temperature Analysis (Distributed Climate Data Science)
For data scientists and climate researchers. You analyze the Berkeley Earth Global Surface Temperature dataset (~65 000 grid cells × 2 000 months) by distributing area-weighted aggregation across 18 AWS Lambda functions using Lithops.
| Experiment | What You Learn | Difficulty |
|---|---|---|
| Global Temperature | Spatial partitioning + area-weighted map-reduce on NetCDF climate data | ⭐⭐⭐⭐ |
The Philosophy
"Don't teach me the theory. Show me the prompt, the code, and the result."
Every experiment includes:
- 🎯 The Prompt — Exactly what we told the AI
- 💻 The Local Code — The script that only works on your machine
- ☁️ The Cloud Code — The distributed version generated by AI
- 📸 The Screenshots — Real execution results, not mockups
- 📐 The Architecture — Mermaid diagrams showing data flow
Ready? Pick an experiment above and let's build something that scales.