Ravin Kumar canyon289

Welcome!

In this workshop, you'll build a lightweight AI agent from scratch —
one that can recognize when it needs external information, call a tool to get it, and continue the conversation based on real results.

You'll see how modern LLMs can:

Detect when a tool call is needed.
Call those tools automatically.
Dynamically adjust their behavior based on new information.

🚀 Getting Started

To run this workshop locally, you'll need to set up Ollama and a Python environment using UV.

1. Setting Up Ollama (Most Critical Step)

We’ll be running Gemma 2B locally with Ollama, so you need to set this up first. This step will require a large download (~10GB total) and some hardware considerations.

Install Ollama

Download and install Ollama from https://ollama.com/.

model.py

Context Manager

https://github.com/pymc-devs/pymc3/blob/master/pymc3/model.py#L277

One of the important pieces of code that handles variable instantion I think.

TODO:

Find where RVs register themselves with this context.
Be able to articulate how the context manager works

	import gym
	import numpy as np
	from collections import deque
	from gym import wrappers

	# Create the Cart-Pole game environment
	env = gym.make('CartPole-v0')
	env = wrappers.Monitor(env, '/tmp/cartpole-experiment-1')

	def relu(x):