Agentic AI Learning Portal

Appearance

Hands-On Projects Guide ​

AI_BOOTCAMP is your dedicated, optimized workspace for building agentic tools and intelligent automations from the ground up during the AI Bootcamp.

  • Linux Path: ~/AI_BOOTCAMP
  • Default Conda Environment: ai_dev (~/miniconda3/envs/ai_dev)

📂 Active Workspace Folder Layout (To Build) ​

As you progress through the Course Syllabus, you will organize ~/AI_BOOTCAMP into the following structured directories:

text
~/AI_BOOTCAMP/
├── requirements.txt (Active Python dependencies)
├── pyproject.toml (Project config & package list)
├── .env (Local api keys & credentials)
├── labs/                  (Per-lesson hands-on labs, Day 1-5)
└── labs/foreman/          (Capstone: The Foreman, see Course Projects)
    ├── owl/               (Research & triage agent, memory/RAG, structured-data schemas, resilient provider calls)
    ├── gnome/              (Execution agent, sandboxing, self-improving skills)
    ├── graph.py            (LangGraph supervisor wiring Owl + Gnome)
    ├── webhook_bridge.py   (FastAPI bridge, real alerts into chat)
    └── dashboard_stream.py (SSE live trace)

See Capstone: The Foreman for the full build.

⚡ Active Shell Commands ​

With your new aliases and lazy-loading shell integrations active, manage your environment instantly:

  1. Jump to this Workspace:
    bash
    cd ~/AI_BOOTCAMP
  2. Activate Primary Conda Environment (ai_dev):
    bash
    conda activate ai_dev
  3. Run claude-code in this folder:
    bash
    npx @anthropic-ai/claude-code

🛠️ Optimizing the ai_dev Conda Environment ​

Your ai_dev environment is extremely well-seeded with libraries like FastAPI, Playwright, Pydantic, SQLAlchemy, and the official google-genai and google-generativeai SDKs.

To keep this environment light and high-performing: ​

  • Use UV for package management: Avoid slow conda resolutions inside ai_dev. Install packages instantly using uv:
    bash
    # Check what is installed
uv pip list
# Install new packages
uv pip install langchain-anthropic langgraph
  • Environment Audits: Clean up build caches occasionally:
    bash
    conda clean --all -y
uv cache clean