Articles
Foundation: Introduction to LangGraph
October 2024
Certificate in LangGraph foundations for building stateful,
multi-actor applications with LLMs.
Training and Finetuning Reranker Models with Sentence
Transformers v4
March 26, 2025
Guide to training and finetuning reranker models with Sentence
Transformers v4.
Mastering RAG: How to Select a Reranking Model
Galileo AI
Guide to selecting reranking models for RAG systems.
Introduction to Recommender Systems: Content-Based,
Collaborative Filtering, and Hybrid Recommendation Engines
Alpha Quantum
Introduction to recommender systems: content-based,
collaborative, and hybrid approaches.
Reverse-mode automatic differentiation from scratch, in Python
June 11, 2020
Building a minimal autodiff framework from scratch with Python
implementation.
A Practical Guide to Contrastive Learning
July 30, 2024
Building SimSiam models with FashionMNIST for self-supervised
learning.
Books
Build a DeepSeek Model (From Scratch)
R. A. Dandekar et al., 2025
Learn how to build DeepSeek's core innovations including
Multi-Head Latent Attention, Mixture-of-Experts, and Multi-Token
Prediction from scratch.
Building Recommendation Systems in Python and JAX
B. Bischof and H. Yee, 2023
O'Reilly Media publication on building recommendation systems
using Python and JAX.
Speech and Language Processing
D. Jurafsky and J. H. Martin, 2024
Comprehensive introduction to NLP, computational linguistics,
and speech recognition.
Research Papers
How Attentive are Graph Attention Networks?
S. Brody et al., 2021
Analysis showing that GAT computes only static attention, and
introduction of GATv2 with dynamic attention that is strictly
more expressive.
LightGCN: Simplifying and Powering Graph Convolution Network
for Recommendation
X. He et al., 2020
A simplified Graph Convolutional Network for collaborative
filtering that removes unnecessary components and achieves
significant performance improvements.
Graph Convolutional Neural Networks for Web-Scale Recommender
Systems
R. Ying et al., 2018
PinSage: A large-scale Graph Convolutional Network deployed at
Pinterest for web-scale recommendation with billions of items
and users.