Skill UI

Covers HuggingFace PEFT workflows for LoRA/QLoRA-based LLM fine-tuning, enabling sub-1% parameter updates on limited GPUs plus adapter serving tips.

This guide provides comprehensive strategies for implementing privacy-safe analytics using PostHog. It addresses critical compliance requirements like GDPR and CCPA, covering explicit consent management (opt-in/opt-out), real-time property sanitization (stripping PII like IP or tokens), and methods for handling Data Subject Access Requests (SAR) and the Right to Erasure. Use this to build highly compliant and privacy-preserving analytics pipelines.

Prediction Monitor Assistant automates machine learning deployment monitoring tasks, offering step-by-step guidance, best-practice patterns, production-ready code and configuration generation, and validation to keep serving and inference workflows reliable.

Auto-activating assistant that guides AWS Sagemaker endpoint deployment, from best practices to production-ready code, configurations, and validation for inference serving and monitoring workflows.

Serve production-grade LLM APIs with vLLM, leveraging PagedAttention, continuous batching, quantization, and tensor parallelism to maximize throughput while keeping latency, GPU usage, and scalability under control for OpenAI-compatible deployments.

SGLang is a high-performance serving framework for LLMs and VLMs that uses RadixAttention prefix caching to deliver structured outputs (JSON/regex/grammar) and agentic workflows with function calls, providing up to 5× faster inference than vLLM across multi-GPU production deployments.

SkyPilot orchestrates ML training, batch jobs, and serving across 20+ clouds with automatic cost-aware region selection, spot-instance recovery, managed clusters, and GPU-aware distributed training, letting you optimize spend and avoid vendor lock-in.

Speeds up LLM inference by combining speculative decoding with Medusa multiple heads and lookahead techniques so chatbots, code assistants, and other real-time workloads can run with 1.5-3.6× lower latency while preserving target-model quality.

Automates guidance for streaming inference setup, covering best practices in ML deployment, serving, monitoring, and production-ready configurations whenever you ask about streaming inference patterns or implementation.

Automates TensorFlow SavedModel deployment support with step-by-step guidance, best-practice patterns, production-ready code/config generation, and validation for serving, monitoring, and optimization scenarios.

Guides TensorFlow Serving setup with step-by-step ML deployment orchestration, enforcing best practices, ready-to-run configurations, and output validation for production inference services.

Optimizes large language model inference on NVIDIA GPUs with TensorRT, delivering 10-100× faster throughput, quantized precision (FP8/INT4), multi-GPU scaling, and serving-ready tooling for production deployments.