Building LLM Agents Made Simple | PyData Boston 2025

Create robust, production-ready LLM applications that actually work—no hype, just practical code you can use tomorrow.

Think about data as connections rather than just points—reveal patterns that traditional analysis misses.

Extract meaningful insights from connected data with practical tools and a new way of thinking about your data.

NumPy with superpowers—automatic differentiation, JIT compilation, and seamless GPU acceleration for lightning-fast scientific computing.

Build intuition by coding your way through probability distributions—where simulation reveals insights that analytical approaches miss.

My take on teaching network analysis and graph theory concepts, using NetworkX. Taught at many conferences since 2015.

From uncertainty to understanding—embrace probabilistic thinking and build models that reflect what you know (and what you don't).

Model, loss, and optimizer: the core components of deep learning. Come learn more; at the end, we even build the beginnings of a deep learning framework!

Bayesian statistical analysis concepts, taught using PyMC3, with Hugo Bowne-Anderson.

Where it all started—thinking about data as networks to reveal hidden patterns in social systems and biological networks.

Think about data as connections rather than just points—reveal patterns that traditional analysis misses.