Skill UI

A comprehensive, low-level Software Development Kit (SDK) for building and managing stateful AI Agents within the Azure AI ecosystem using .NET. This SDK provides core clients for managing agent lifecycles, persistent conversation threads, message history, complex run execution, tool/function calling, and integrating file search through vector stores. It is essential for developing enterprise-grade, multi-turn conversational AI applications.

A comprehensive, low-level Software Development Kit (SDK) for building and managing stateful AI Agents within the Azure AI ecosystem using .NET. This SDK provides core clients for managing agent lifecycles, persistent conversation threads, message history, complex run execution, tool/function calling, and integrating file search through vector stores. It is essential for developing enterprise-grade, multi-turn conversational AI applications.

Gastown is a robust multi-agent orchestration system designed for Claude Code. It enables the simultaneous execution of multiple AI agents on complex projects. It manages the entire development lifecycle, from setting up workspaces and tracking work items (beads) to managing agent lifecycles (polecats/crew) and merging final code (Refinery). Use it when coordinating complex, parallel AI tasks.

Execute complex and compute-intensive workloads directly on fully managed Hugging Face cloud infrastructure. This skill eliminates the need for local setup, allowing users to run jobs using cloud CPUs, GPUs, or TPUs. Ideal for large-scale data processing, batch inference, reproducible model training, and scheduled automation tasks, while persisting results to the Hugging Face Hub.

Fully managed, auto-scaling vector database tailored for production AI workloads such as RAG, recommendations, and semantic search, offering hybrid dense+sparse search, namespaces, metadata filters, and <100ms latency on serverless infrastructure.

Use Azure OpenAI Realtime Mini to turn text prompts into streamed PCM audio, convert it to WAV, and deliver base64 payloads for frontend playback, handling transcripts and voice tuning for podcast-style narration.

Sentence Transformers provides 5000+ pre-trained models that generate sentence, text, and image embeddings for RAG, semantic search, clustering, and multilingual production workloads without relying on external APIs.

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.

A comprehensive guide demonstrating the full lifecycle of GPU resource utilization on Vast.ai. This includes searching for available hardware using CLI and REST APIs, provisioning a dedicated GPU instance (e.g., using PyTorch images), connecting via SSH, running deep learning workloads, and finally, securely destroying the instance to stop billing. Ideal for testing cloud integration or managing ML compute pipelines.

A comprehensive checklist for deploying large-scale, production-ready GPU workloads on Vast.ai. It covers critical lifecycle stages including account authentication, instance validation, data encryption, persistent checkpointing strategies, fault tolerance for spot instances, robust monitoring (utilization & cost), and strict cost controls, ensuring reliable and stable ML/AI deployments.

Modal delivers cloud-native serverless Python with on-demand GPUs, persistent storage, and web endpoints to deploy ML workloads, batch jobs, and APIs without writing Dockerfiles.

This infrastructure provides a robust solution for running powerful AI agents in remote, sandboxed environments. It overcomes the limitations of local execution by ensuring unlimited concurrency, consistent environments, and collaborative features. Use this to build complex background coding agents, multi-client interfaces (e.g., Slack, Web), or scale agent workloads far beyond local machine constraints.