Free AI library

Technical guidesFree forever.

177 plain-English guides, structured learning paths, and an open library — built by an independent 501(c)(3) nonprofit so anyone can understand modern AI.

177Free guides
1Topic tracks
~2 minPer guide
~6hReading time

Start here

Five outcome-based courses

Each course includes explicit outcomes, mapped competencies, practice activities, and an applied capstone.

Topic tracks

Browse by track

Jump into the area you care about. Every track has multiple plain-English guides.

Full library

All guides

177 of 1019 guides shown. Filter by track or search above.

Technical

Depthwise Separable Convolutions

Depthwise separable convolutions factor a standard convolution into two cheaper steps, slashing the number of multiplications and parameters.

2 min readRead
Technical

Dilated and Atrous Convolutions

Dilated convolutions (also called atrous convolutions) insert gaps between filter weights so a kernel covers a much larger area without adding parameters.

2 min readRead
Technical

Highway Networks and Skip Connections

Skip connections let information jump past layers, and highway networks were an early gated version of this idea.

2 min readRead
Technical

Domain Adaptation

Domain adaptation is a set of techniques for making a model trained on one kind of data (the source domain) work well on a different but related kind of data…

2 min readRead
Technical

InfiniBand and RDMA Networking

InfiniBand is a high-speed, low-latency interconnect that links servers and GPUs in AI clusters, and RDMA lets one machine read or write another's memory…

2 min readRead
Technical

Collective Communication and NCCL

Collective communication is how a group of GPUs exchange and combine data, and NCCL is NVIDIA's library that makes those exchanges blazingly fast.

2 min readRead
Technical

Feature Engineering Pipelines and Data Versioning

Feature engineering pipelines transform raw data into the numerical signals models actually learn from, while data versioning tracks exactly which data…

2 min readRead
Technical

Online and Offline Feature Serving Skew

Training/serving skew happens when the features a model learns from offline differ from the features it actually receives in production, quietly wrecking…

2 min readRead
Technical

Kubeflow and ML Pipeline Orchestration

Kubeflow is an open-source toolkit that runs machine learning workflows on Kubernetes, turning model training and deployment into reproducible, containerized…

2 min readRead
Technical

Apache Airflow for ML Workflows

Apache Airflow is an open-source platform for authoring, scheduling, and monitoring workflows as code.

2 min readRead
Technical

MLflow and Model Lifecycle Tracking

MLflow is an open-source platform for managing the machine learning lifecycle, from experiment tracking to model packaging and deployment.

2 min readRead
Technical

Expert Parallelism for MoE Serving

Expert parallelism splits a Mixture-of-Experts model's many feed-forward 'experts' across different GPUs so each device holds only a slice of the parameters.

2 min readRead

Finished reading? Prove it.

Check what you learned with topic quizzes, then explore our structured courses and current certification requirements. Core guides remain free to read.