meta-skills docs

Agent-facing documentation pages, served as raw markdown. Start at llms.txt; source lives at ryan-minato/meta-skills.

• 3-D Vision & Neural Rendering — Point-cloud and mesh processing, differentiable 3D ops, NeRF and Gaussian-splatting tooling.
• Computer Vision — Core vision libraries, pretrained-model zoos, augmentation, and detection/segmentation/tracking tooling.
• DataFrames, Storage & Data Quality — Dataframe and embedded-SQL libraries, columnar and scientific storage formats, parallel and object-storage I/O, and schema/quality validation.
• Deep Learning Frameworks — General DL and tensor frameworks — PyTorch, TensorFlow, Keras, JAX and its ecosystem, MLX — and their documentation roots.
• Distributed & GPU Data Compute — Dask, cluster analytics engines (Spark, Flink, Trino), the Ray ecosystem, and RAPIDS GPU data science.
• Geospatial Analysis — Vector and raster stacks, CRS handling, and spatial engines for geographic data.
• GPU Kernels, Compilers & Autodiff — JIT compilers, GPU kernel languages and toolchains, kernel-level libraries, and automatic differentiation of programs.
• Graphs & Graph Learning — Network analysis, graph neural networks, knowledge-graph embeddings, and probabilistic graphical models.
• HPC, MPI & Cluster Scheduling — MPI stacks, cluster schedulers, scientific workflow managers, and GPU/multi-node collective communication.
• Hugging Face Ecosystem — Everything documented under the Hugging Face docs root plus HF-owned products — model, data, training, optimization, serving, and app libraries.
• Image & Video Generation — Diffusion-model generation UIs, LoRA trainers, conditioning adapters, and open video-generation projects.
• Inference Engines & Model Serving — LLM inference engines, local runtimes, and model-serving platforms for running models locally or behind an API.
• LLM Applications, RAG & Agents — Orchestration frameworks, agent SDKs, structured output, gateways, and guardrails for building LLM applications.
• LLM Evaluation & Observability — Benchmark harnesses for scoring model, RAG, and agent quality, plus tracing platforms for LLM calls in development and production.
• Quantization, Model Compilers & Edge Runtimes — Quantization and compression toolkits plus model compilers and cross-platform runtimes for compiled, mobile, and edge deployment.
• Numerical Computing & Statistics — Scientific-computing platforms, array and statistics libraries, and the math kernels and sparse solvers C/C++/Julia numerics link against.
• OCR & Document AI — OCR engines, document layout analysis, and PDF parsing tooling for reading text from images and documents.
• Optimization, Differential Equations & FEM — Mathematical-programming solvers, ODE/DAE integrators, and PDE/FEM/multiphysics frameworks.
• Pipelines, Orchestration & MLOps — Workflow orchestrators, analytics engineering, experiment tracking and versioning, and ML/production monitoring.
• Probabilistic Programming & Causal Inference — Bayesian modeling and MCMC/variational inference libraries, plus causal effect estimation and discovery.
• Recommender Systems — Ranking, retrieval, and CTR frameworks for recommendation and candidate-generation systems.
• Reinforcement Learning — RL algorithm and training frameworks, plus the environments and simulators agents train against.
• Scientific & Domain ML — ML libraries for medicine, biology, chemistry, materials, and physics-informed modeling.
• Speech, Voice & Audio — Speech recognition and diarization, TTS and voice cloning, voice conversion and separation, and music/generative-audio tooling.
• Tabular ML, AutoML & Interpretability — Classic ML and gradient boosting, model interpretability, and AutoML and hyperparameter-optimization libraries.
• Time Series & Forecasting — Forecasting, feature-extraction, and change-point detection libraries for time-series data.
• Distributed Training & Finetuning — Distributed-training stacks and finetuning frameworks — DeepSpeed, Megatron, Lightning, NeMo, torchtune, Axolotl — for multi-GPU, multi-node, and post-training work.
• Privacy, Robustness & Federated Learning — Differential privacy, adversarial robustness, fairness auditing, PII redaction, interpretability, and federated-learning frameworks.
• Vector Search & Databases — Local ANN index libraries, vector databases, and search engines for embedding retrieval.
• Visualization, Data Apps & Notebooks — Plotting libraries, dashboards and small data apps, scientific 3D rendering, and notebook or publishing tooling.