Readings

PySpark vs Pandas: Performance, Memory Consumption and Use Cases

Parallel/distributed computing advantages and when to use Spark versus Pandas.

Distribution Styles — Amazon Redshift

Three distribution key types for data clustering in Redshift.

Evaluating the Query Plan — Amazon Redshift

Sort key determination and join efficiency optimization.

Spark vs Hadoop: MapReduce, Performance, and Resource Management

Comparative analysis of distributed computing frameworks.

Attention Is All You Need

The foundational transformer architecture paper underlying modern AI models.

Deep Double Descent: Where Bigger Models and More Data Hurt

Exploring the curse of dimensionality and learning curve analysis for data sufficiency.

KAN: Kolmogorov-Arnold Networks

New neural network architecture offering faster training for smaller networks.

Mamba: Linear-Time Sequence Modeling with Selective State Spaces

Alternative to Transformers for efficient long-sequence processing.

Feast: Bridging ML Models and Data

Open-source feature store for production ML.

Exploring Bayesian Optimization

Hyperparameter tuning using Bayesian techniques and Thompson Sampling.

Dimensionality Reduction Algorithms: Strengths and Weaknesses

Feature Extraction vs Feature Engineering

The Case for Sparsity in Neural Networks, Part 1: Pruning

Machine Learning in Production: Why You Should Care About Data and Concept Drift

Model degradation from data/concept drift in production environments.

Monitoring Machine Learning Models in Production

Key oversight considerations for long-term model performance management.

Baselines in Data Analysis

Setting human-level performance baselines for realistic ML targets.

Machine Learning: The High-Interest Credit Card of Technical Debt

Tech debt issues including unstable data dependencies and pipeline complexity.