Skill UI

This skill provides comprehensive guidance for building production-grade MLOps pipelines. It covers the entire machine learning lifecycle, from initial data ingestion and rigorous feature engineering to model training, advanced validation (A/B testing, performance drift), and robust deployment strategies (Canary, Blue-green). It enables the design and orchestration of scalable, reproducible, and observable ML systems using tools like Airflow and Kubeflow.

nanoGPT is a simplified, minimalist implementation designed specifically for educational purposes. It reproduces the core architecture of GPT-2 (124M) using clean, hackable code, allowing users to understand the entire transformer pipeline from scratch. The framework supports full workflows, including data preparation on various datasets (e.g., Shakespeare, OpenWebText), model training, and text generation, making it ideal for students and researchers learning NLP and deep learning principles.

Phoenix is an open-source observability platform for tracing, evaluating, and monitoring LLM applications; use it to debug prompts, run dataset evaluations, build experiment pipelines, and watch production inference in real time.

Provides Meta AI’s Segment Anything Model for zero-shot image segmentation via interactive prompts or automatic mask generation, enabling rapid dataset creation and integration in vision pipelines.

Retrieve AlphaFold’s AI-predicted protein models via UniProt IDs, download CIF/PDB coordinates, inspect confidence metrics, and feed them into structural analysis or drug-discovery pipelines.

PyHealth delivers a full-stack Python toolkit for building, evaluating, and deploying clinical machine learning models, covering EHR datasets, medical coding systems, physiological signals, and healthcare-tailored architectures such as RETAIN, SafeDrug, and Transformer-based pipelines.

A comprehensive skill for leveraging massive pre-trained transformer models across diverse AI domains, including Natural Language Processing (NLP), Computer Vision (CV), and Audio. Use this skill for advanced tasks such as text generation, classification, question answering, summarization, object detection, and model fine-tuning on proprietary datasets. It provides streamlined pipelines and granular control for deployment.

A framework for building production-ready, low-latency, multimodal AI agents. It allows developers to orchestrate complex pipelines that seamlessly handle real-time audio, video, and text data. Key capabilities include integrating multiple AI services (LLMs, STT, TTS), managing conversation context, and enabling tool use and function calling for advanced, interactive AI applications.

Modal is a serverless cloud platform designed for executing Python-based AI/ML workloads. It provides on-demand access to powerful GPUs (e.g., H100, A100), scalable serverless functions, and custom containerization. Use cases include deploying ML models, running large-scale training/inference jobs, creating web APIs, and scheduling data pipelines, all defined entirely in code.

This guide provides comprehensive, version-aware documentation for using PufferLib in reinforcement learning pipelines. It details best practices for adapting Gymnasium/PettingZoo environments, managing different PufferLib versions (3.0.0 vs 4.0 source), and executing safe, reproducible training and evaluation routines. Key focus areas include environment contract validation, sandboxing native builds, and maintaining data provenance.

PyHealth is a modular Python toolkit designed for developing deep learning pipelines in clinical and healthcare settings. It provides a standardized 5-stage framework (Dataset → Task → Model → Trainer → Metrics) to handle complex, multi-modal data, including EHR records, physiological signals, and medical imaging. Use cases include predicting mortality, readmission risks, recommending drugs, or performing medical code mapping (ICD/ATC).

PyTorch Lightning is an advanced deep learning framework built on PyTorch. It simplifies the entire ML workflow, allowing users to organize models into LightningModules, automate training loops using the Trainer, and manage complex data pipelines with DataModules. It supports advanced features like multi-GPU/TPU scaling (DDP, FSDP), integrated logging (W&B, MLflow), and custom callbacks, making large-scale neural network development efficient and reproducible.