Getting Started

Get pAI MSc running in under 10 minutes.

Prerequisites

• Python 3.10+ (3.11 recommended)
• Git
• An API key for at least one LLM provider (Anthropic, OpenAI, Google, or DeepSeek)
• (Optional) LaTeX for PDF output
• (Optional) SLURM for HPC cluster execution

Step 1: Install

git clone https://github.com/PoggioAI/PoggioAI_MSc.git
cd PoggioAI_MSc
git checkout MSc_Prod

Check your Python version:

python3 --version   # Must be 3.10 or higher

# venv
python3 -m venv .venv
source .venv/bin/activate
pip install --upgrade pip
pip install -e .

# conda
conda create -n msc python=3.12 -y
conda activate msc
pip install -e .

Optional extras:

pip install -e ".[all]"          # All optional dependencies
pip install -e ".[web]"          # Web search and retrieval
pip install -e ".[experiment]"   # PyTorch, transformers, datasets
pip install -e ".[docs]"         # Document conversion tools

Step 2: Setup Wizard

msc setup

The setup wizard will:

1. Detect your platform (OS, GPUs, SLURM availability, installed tools)
2. Configure API keys, saved in ~/.msc/.env
3. Select models and optional counsel models
4. Set budget limits
5. Configure notifications (optional)

Step 3: Verify Installation

msc doctor

Step 4: Run Your First Paper

msc run "What are the key differences between transformer and state-space models?" --preset quick

Step 5: Check Results

msc status
msc logs -f
msc runs

Output manuscripts are saved to results/ by default.

What gets produced

Next Steps

• Configuration
• Usage
• HPC / SLURM Setup