Discover the Best AI Resources

roadmap.sh (Developer Roadmap)

Interactive, community-driven learning roadmaps and guides that map skills, technologies, and curated resources for developer career paths — covering frontend, backend, DevOps, ML/AI, MLOps, prompt engineering and more. Clickable nodes link to tutorials, best practices and question banks to guide study and hiring prep.

Recasts a scatter of competing graph-network designs as one message-passing recipe — propagate, aggregate, read out — then proves it on QM9, hitting chemical accuracy on most molecular property targets without hand-built descriptors.

Isolates relational reasoning into a tiny plug-in module that scores pairwise object relations, bolting onto CNN/LSTM encoders to hit super-human 95.5% on CLEVR — and proving plain convnets lack this capacity on their own.

The 2017 paper that replaced recurrence with pure self-attention, making sequence models fully parallelizable — and, almost as a side effect, laying the architectural foundation for nearly every large language model that followed, from BERT to GPT.

Weights & Biases

Tracks every ML run — hyperparameters, metrics, checkpoints, dataset versions — into one dashboard you share as a live report, with Sweeps for tuning and a model registry. Weave extends it to LLM apps: tracing, evals, and production monitoring.

CatBoost

Trains gradient-boosted decision trees with native categorical-feature handling, GPU acceleration, and production-ready prediction APIs. A strong fit for tabular ML when preprocessing categories into numeric features would add noise or leakage.

Nx

Manages polyglot monorepos by caching unchanged outputs and running only affected tasks. Built with Rust and extensible in TypeScript; includes integrated CI features (remote caching, task distribution) and AI-native tooling such as a CLI optimized for autonomous agents and self-healing CI.

fairseq

Sequence modeling toolkit for training custom models for translation, summarization, and language modeling. Reference implementation behind RoBERTa, BART, mBART, XLM-R, and wav2vec 2.0, with multi-GPU and mixed-precision training.

great_expectations

Expresses data quality checks as reusable, declarative "expectations" and auto-generates human-readable validation reports and docs; integrates with Python data stacks to enforce and monitor data reliability in ML and analytics pipelines.

ONNX

Defines a portable model format and operator set for moving trained machine learning models across frameworks, runtimes, and hardware targets without locking the model to one toolchain.

PyTorch Geometric (PyG)

Provides reusable PyTorch modules and utilities to build, train and scale Graph Neural Networks — includes many implemented GNN layers, benchmark datasets, minibatch/sampling loaders, and support for large-scale, heterogeneous, temporal and point-cloud graphs.

Ray (by Anyscale)

Scales any Python or ML workload across CPUs and GPUs with a few decorators, instead of rewriting code for Spark or MPI. Bundles libraries for distributed training, hyperparameter tuning, RL, batch inference, and online model serving on one cluster.