AI & Machine Learning Bundle

Prompt engineering, LLM integration, agent design, model fine-tuning, and AI workflow skills for ML practitioners.

Included Skills (90)

  1. Prompt Engineering Specialist — This skill helps developers optimize prompts, validate agent workflows, and measure RAG performance using data-driven evaluation techniques to ensure reliable, model-agnostic outputs.
  2. Model Tuning Assistant — This tool enables developers to fine-tune open and Gemini models within the Agent Platform infrastructure to optimize performance for specific application requirements.
  3. AI Security Assessor — Assess AI/ML systems for vulnerabilities like prompt injection, model inversion, and data poisoning, aiding security engineers and AI developers.
  4. LangChain Application Builder — Build LLM-powered applications using LangChain's framework for chains, agents, RAG, tool integration, memory, and deployment.
  5. RAG Security Auditor — Probes Retrieval-Augmented Generation pipelines for indirect prompt injection vulnerabilities to help security engineers validate guardrails and secure document-based AI systems.
  6. LLM Training Tool — Train a small GPT model from scratch, helping users understand LLM architecture and build custom language models.
  7. mastra — Mastra is a TypeScript framework that lets developers build production‑ready AI agents, RAG pipelines, and complex workflows with type‑safe definitions and reliable tool integration. It provides vector‑based knowledge retrieval, multi‑step branching, and robust error handling while connecting to more than fifty third‑party services. TypeScript teams can create sophisticated agent capabilities without relying on Python libraries.
  8. LLM Prompt Engineer — This skill helps AI developers design effective prompts for LLMs, improving output quality and reliability across various applications.
  9. Fireworks AI Assistant — Provides expert guidance for Fireworks AI, helping developers integrate the inference API, fine-tune models, and deploy custom endpoints.
  10. Together AI Platform — Together AI provides a cloud platform for developers to run and fine-tune open-source AI models via inference APIs.
  11. Mask2Former Segmentation Trainer — This tool enables developers to train, evaluate, and deploy high-quality universal image segmentation models using the NVIDIA TAO Mask2Former transformer-based architecture.
  12. Gemini Enterprise Integration — This skill provides developers with comprehensive guidance for integrating Gemini models into enterprise applications using the Google Gen AI SDK and Agent Platform.
  13. Agent Tool Security — Implements defense-in-depth controls for AI agents to prevent unauthorized tool execution and mitigate risks from prompt injection or malicious tool usage.
  14. AI Eval CI — Automate AI agent and LLM evaluations in CI/CD pipelines, ensuring quality and preventing regressions before deployment, benefiting developers and AI engineers.
  15. Mistral AI Interface — Provides access to Mistral AI's language models for code generation, multilingual tasks, and GDPR-compliant AI inference.
  16. Hugging Face Toolkit — Enables users to leverage Hugging Face's libraries for tasks like model inference, fine-tuning, and publishing, benefiting ML practitioners.
  17. LLM Security Guardrails — This skill implements robust input and output validation guardrails for LLM applications to ensure safety, compliance, and protection against malicious prompt injections.
  18. LLM Observability Proxy — Helicone logs LLM requests, enables caching/rate-limiting, and provides cost analytics, benefiting developers using OpenAI, Anthropic, and other LLM providers.
  19. AI Observability Expert — Assists developers in monitoring LLM applications using Arize and Phoenix for tracing, evaluation, and performance analysis.
  20. TAO Image Classifier — This tool enables developers to train, distill, and quantize PyTorch-based image classification models using NVIDIA TAO for optimized deployment on edge hardware.
  21. RAG Engine Manager — This tool enables developers to manage and query Agent Platform RAG Engine corpora and retrieve grounded contexts using the Google GenAI SDK.
  22. GKE RAG Architect — This skill provides architectural guidance and deployment strategies for building secure, high-accuracy RAG-enabled enterprise search solutions using GKE and vector-enabled SQL databases.
  23. Vertex AI Gemini — This skill helps enterprises deploy and fine-tune Gemini models on Google Cloud Vertex AI with enterprise-grade features.
  24. Agent Factory — Claude Code agent generation system that creates custom agents and sub-agents with enhanced YAML frontmatter, tool access patterns, and MCP integration support following proven production patterns
  25. Persistent Agent Memory — Adds file-based, vector, and semantic memory to AI coding agents, enabling them to retain context between sessions and learn over time.
  26. Serverless AI Inference — Deploy and run AI models instantly with fal.ai, enabling developers to generate media and run ML inference at scale.
  27. Modal Serverless GPU — Deploys and runs Python functions on cloud GPUs, enabling ML model deployment and inference without infrastructure management for developers.
  28. ONNX Model Converter — Facilitates model interoperability by converting models to ONNX format, enabling optimization and cross-platform deployment for AI developers.
  29. Groq API Assistant — Provides expert guidance for integrating Groq's ultra-fast LLM inference API into real-time AI applications, assisting developers.
  30. Claude Code Orchestrator — Orchestrates multi-agent Claude Code teams, assigning roles, coordinating tasks, and managing workflows for complex AI development projects.
  31. Crawl4AI Web Crawler — Assists developers in extracting clean, structured website data for AI applications like LLM training and RAG pipelines using Crawl4AI.
  32. LLM Observability Expert — Assists developers in instrumenting AI pipelines with automatic tracing using Traceloop's OpenLLMetry SDK for enhanced observability.
  33. Pinecone Vector Database — This skill helps AI developers use Pinecone, a managed vector database, to build semantic search and RAG applications.
  34. SegFormer Training Assistant — This tool enables developers to train, evaluate, and deploy lightweight SegFormer models for efficient real-time semantic segmentation tasks using NVIDIA TAO.
  35. Model Deployment Manager — This tool enables developers to deploy, monitor, and manage open or custom model resources within the Agent Platform infrastructure efficiently.
  36. BEVFusion Training Assistant — This tool enables autonomous driving developers to train, evaluate, and run inference for multi-sensor 3D object detection models using BEVFusion and TAO.
  37. LangGraph Agent Builder — Build stateful, multi-step AI agents and workflows with LangGraph, assisting developers in creating complex AI applications.
  38. TAO DINO Detector — This skill enables developers to train, evaluate, and deploy DINO transformer-based 2D object detection models using the NVIDIA TAO toolkit workflow.
  39. Triton Inference Server — Deploys AI models at scale, supporting multiple frameworks and hardware, benefiting data scientists and machine learning engineers.
  40. Hugging Face Integrator — This skill enables developers to load, perform inference, and fine-tune pre-trained models from the Hugging Face Hub for diverse multimodal machine learning tasks.
  41. Unified LLM Interface — LiteLLM provides a single API to call 100+ LLMs, enabling developers to easily switch providers and manage LLM access.
  42. LLM Observability Expert — Assists developers in tracing and monitoring LLM applications using Langtrace, an open-source observability platform built on OpenTelemetry.
  43. TensorFlow Expert Assistant — Assists developers in building, training, and deploying neural networks using the TensorFlow machine learning framework and its related tools.
  44. Multi-Agent Orchestrator — This protocol enables multiple AI agents to collaborate, delegate tasks, and share context, helping developers build complex, multi-agent workflows with structured communication.
  45. AI Agent Alerting — This tool generates Terraform configurations for OpenTelemetry-based alerting policies to help developers monitor AI agent latency, error rates, token usage, and quality.
  46. Sparse4D Perception Trainer — This tool enables developers to train, evaluate, and deploy Sparse4D models for multi-camera temporal 3D object detection and tracking tasks.
  47. Gemini Agent Manager — This tool enables developers to programmatically provision, configure, and manage stateful Gemini Enterprise Agent resources, including custom skills, tools, and sandboxed files.
  48. Panoptic 3D Reconstruction — This skill enables developers to train and deploy NVPanoptix3D models for high-fidelity 3D scene segmentation and occupancy completion from posed RGB images.
  49. Mask Grounding Trainer — This tool enables developers to train, evaluate, and deploy open-vocabulary instance segmentation models using text-prompted Mask Grounding DINO workflows within NVIDIA TAO.
  50. Video Action Classifier — This tool enables developers to train, evaluate, and deploy NVIDIA TAO action recognition models using RGB, optical flow, or multi-stream video inputs.
  51. Cosmos Embedder Tool — This tool enables developers to fine-tune, evaluate, and perform inference with Cosmos-Embed1 models for advanced video-text retrieval and semantic video analysis tasks.
  52. OneFormer Segmentation Trainer — This tool enables developers to train, evaluate, and deploy universal image segmentation models using the TAO OneFormer architecture for panoptic, instance, and semantic tasks.
  53. GKE JobSet Troubleshooter — This tool autonomously diagnoses and resolves GKE JobSet interruptions for AI/ML workloads, helping engineers quickly identify root causes like preemptions and node failures.
  54. ML Integrity Auditor — This tool identifies poisoned training data and backdoored machine learning models to help security engineers ensure the integrity of their AI supply chain.
  55. Multi-Agent Orchestrator — Coordinate complex multi-agent systems by managing agent communication, task routing, memory sharing, and quality control to improve performance in autonomous AI workflows.
  56. LLM Prompt Tester — This skill helps developers systematically test and evaluate LLM prompts using the open-source Promptfoo framework for prompt engineering.
  57. Genkit Development Assistant — Build and debug AI-powered applications using Genkit in Node.js and TypeScript to streamline agent creation, flow management, and prompt engineering for developers.
  58. Extensible AI Agent — Goose is an open-source AI agent that installs software, executes commands, edits files, runs tests, and manages infrastructure, benefiting developers and DevOps engineers.
  59. PEFT Fine-Tuning — Fine-tune LLMs efficiently using LoRA, QLoRA, and other PEFT methods, helping users adapt models to specific tasks on consumer GPUs.
  60. OpenAI API Integration — Integrate OpenAI's powerful APIs like GPT and DALL-E into applications for text generation, image creation, and more.
  61. Replicate Model Runner — Run and fine-tune open-source machine learning models via API, assisting developers and researchers with model deployment and experimentation.
  62. PyTorch Lightning Assistant — This skill helps deep learning researchers and engineers streamline neural network training by automating boilerplate code and managing complex multi-device distributed workflows.
  63. OpenAI Agents Expert — Assists developers in building production-grade AI agents using the OpenAI Agents SDK with tool calling and safety controls.
  64. OpenRouter API Expert — Assists developers in leveraging OpenRouter's unified API to access and manage diverse LLMs for multi-model strategies.
  65. Haystack Pipeline Builder — This skill helps developers build production-ready RAG pipelines and LLM applications using the Haystack open-source framework.
  66. Sequence Packing Optimizer — This tool enables efficient sequence packing and long-context training configurations in Megatron-Bridge to help developers optimize LLM and VLM performance during model training.
  67. Megatron Recipe Optimizer — This tool helps AI engineers select and customize optimal Megatron Bridge training and benchmarking recipes based on hardware, model, and performance requirements.
  68. Nemotron Safety Architect — This tool helps developers create custom safety policies, taxonomy, and inference prompts for NVIDIA Nemotron content-safety guardrails to ensure robust model governance.
  69. Multi-Agent Orchestration — Build and orchestrate multi-agent AI systems using Microsoft's Agent Framework for defining agents and workflows in Python or .NET.
  70. Opus Migration Assistant — Migrates code and prompts from older Claude models to Opus 4.5, updating model strings and adjusting prompts for behavioral differences.
  71. AI Safety Guardrails — Implement safety guardrails for AI systems, including content filtering and prompt injection detection, to ensure responsible AI practices.
  72. LLM Tool Calling — Enables LLMs to interact with external tools and APIs, empowering developers to build AI agents with real-world action capabilities.
  73. LLM Observability Tool — LangSmith helps developers monitor, trace, debug, and evaluate their LLM applications to improve performance and quality.
  74. Automated LLM Red-Teaming — This tool integrates Promptfoo and DeepTeam into CI/CD pipelines to automate adversarial security testing for LLM applications against critical industry vulnerability benchmarks.
  75. Cerebras Inference Guide — Provides expert guidance for developers integrating Cerebras Inference, an ultra-fast LLM inference service, into applications needing rapid token generation.
  76. Strategic AI Advisor — This tool provides startup founders and CAIOs with strategic guidance on model selection, regulatory compliance, cost economics, and organizational AI team scaling.
  77. RouterBase Integration Gateway — This skill enables developers to seamlessly migrate OpenAI-compatible applications to RouterBase for advanced model routing, fallback management, and multi-modal media generation capabilities.
  78. Embedded Vector Database — LanceDB enables serverless, zero-config vector search for AI applications, offering table creation, vector search, and multimodal embeddings locally.
  79. Postgres Vector Search — Enables storing and searching vector embeddings directly within PostgreSQL, ideal for developers seeking similarity search and RAG capabilities without separate vector databases.
  80. AI SDK Assistant — This assistant helps developers build AI-powered applications by providing accurate, up-to-date guidance on Vercel AI SDK features, integrations, and best practices.
  81. ML Lifecycle Manager — MLflow manages the ML lifecycle, enabling experiment tracking, model registry, and deployment for data scientists and ML engineers.
  82. Weave AI Tracker — This skill helps AI developers trace LLM calls, evaluate outputs, and debug AI pipelines using the Weave toolkit.
  83. LLM Observability Expert — Assists developers using Langfuse, the open-source LLM engineering platform, to trace calls, evaluate quality, and manage prompts for AI applications.
  84. RAG Pipeline Architect — This tool helps developers design, optimize, and evaluate production-grade RAG pipelines by utilizing data-driven chunking strategies and rigorous retrieval performance metrics.
  85. Workflow Script Generator — This skill designs and writes multi-agent workflow scripts for Claude Code, helping users automate repeatable, multi-step tasks.
  86. ChromaDB Vector Database — Assists developers with storing, searching, and managing vector embeddings using ChromaDB for RAG pipelines and semantic search.
  87. LlamaIndex RAG Assistant — Helps build retrieval-augmented generation (RAG) pipelines and knowledge assistants using the LlamaIndex framework for data-augmented LLM applications.
  88. PydanticAI Agent Expert — Assists developers in building type-safe AI agents using PydanticAI, a Python framework for structured outputs and tool definitions.
  89. Agent Development Assistant — Provides guidance and best practices for developing agents, including structure, system prompts, and triggering conditions, specifically for Claude Code plugins.
  90. OCDNet Text Detector — This tool enables developers to train, evaluate, and deploy OCDNet models for detecting arbitrary-oriented text regions within complex natural images.