Skill UI

An expert guide and framework for building production-grade, sophisticated AI agent systems. This skill specializes in utilizing the latest LangChain 0.1+ and LangGraph APIs to implement multi-agent orchestration, advanced Retrieval-Augmented Generation (RAG) pipelines (e.g., RAG Fusion, HyDE), and complex state management. It covers comprehensive topics from architectural design (Plan-and-Execute, ReAct Agents) to production deployment (FastAPI streaming, LangSmith observability, caching, and robust error handling).

Extends transformer context windows using RoPE, YaRN, ALiBi, and position interpolation so long documents (32k-128k+ tokens) can be processed, pre-trained models lengthened, and efficient positional encodings or extrapolation strategies deployed.

This skill guides the design and implementation of a complete, production-grade Machine Learning pipeline. It employs a multi-agent orchestration workflow, systematically covering all stages of the ML lifecycle: data ingestion and quality assurance, feature engineering, model training, optimization, and robust deployment (MLOps). It ensures built-in reproducibility, monitoring, and scalability.

Automate the validation and testing of prompts and AI responses within your CI/CD pipeline using Mistral AI. This integration provides robust prompt regression testing, model quality checks, automatic cost estimation for API usage, and deployment gates, ensuring the stability and reliability of your AI features upon every pull request.

Guides researchers through drafting publication-ready machine learning papers for NeurIPS, ICML, ICLR, ACL, AAAI, and COLM by combining expert writing philosophy with LaTeX templates, citation verification APIs, and collaborative workflows.

Automated guidance for model explainability tooling within ML training workflows, giving step-by-step best practices, production-ready code, and validation for experiments whenever users mention explainability tool needs.

Combines fine-tuned models with Mergekit to blend math, coding, and chat expertise without retraining, covering SLERP, TIES, DARE, Task Arithmetic, and linear merges for rapid experimentation and deployment.

Guides researchers through nnsight’s tracing and activation APIs so you can inspect or modify PyTorch internals locally or via NDIF-powered remote runs on 70B+ foundation models without local GPUs.

This guide provides comprehensive steps to install and authenticate the OpenEvidence SDK/API. It covers setting up environment variables, installing required packages (supporting both Node.js and Python), and verifying the connection using sample code. This is essential for developers integrating OpenEvidence's medical AI capabilities for clinical decision support and evidence-based queries into a new application.

High-performance RLHF framework that combines Ray, vLLM, and ZeRO-3 to accelerate PPO/GRPO/RLOO/DPO training for 7B-70B+ models on distributed GPU clusters, providing a single-stack workflow for reward models and policy optimization.

Use this tool to query, filter, and compare over 400 available AI models from OpenRouter. It allows users to check pricing, context lengths, model capabilities (e.g., tool calling), and select the best model based on specific needs, ensuring optimal performance and cost efficiency in AI development.

A comprehensive toolkit designed to manage the complex data flow associated with Perplexity search workflows. This skill implements critical security measures, including PII sanitization in user queries, validates external citations for accuracy, optimizes performance via result caching with freshness policies, and manages conversation context to maintain coherence and reliability. Essential for building robust, compliance-aware AI applications.