# NVIDIA NeMo Microservices

> A modular set of tools to customize, evaluate, and secure large language models (LLMs) while optimizing AI applications across Kubernetes clusters

NVIDIA NeMo Microservices provides a comprehensive platform for the entire LLM lifecycle, including fine-tuning, evaluation, inference, and guardrails implementation. It supports both on-premises and cloud-based Kubernetes deployments.

## Core Microservices

- [NeMo Customizer](https://docs.nvidia.com/nemo/microservices/latest/fine-tune/index.html.md): Fine-tune LLMs using supervised and parameter-efficient techniques
- [NeMo Evaluator](https://docs.nvidia.com/nemo/microservices/latest/evaluate/index.html.md): Comprehensive evaluation capabilities for LLMs with benchmarks and custom evaluations
- [NeMo Guardrails](https://docs.nvidia.com/nemo/microservices/latest/guardrails/index.html.md): Safety checks and content moderation for LLM endpoints

## Platform Components

- [Entity Store](https://docs.nvidia.com/nemo/microservices/latest/api/entity-store.html.md): Tools to manage entities like namespaces, projects, datasets, and models
- [Deployment Management](https://docs.nvidia.com/nemo/microservices/latest/api/deployment-management.html.md): API to deploy NIM on Kubernetes clusters
- [NIM Proxy](https://docs.nvidia.com/nemo/microservices/latest/api/nim-proxy.html.md): Unified endpoint for accessing deployed NIM for inference tasks

## Documentation

- [Getting Started](https://docs.nvidia.com/nemo/microservices/latest/get-started/index.html.md): Initial setup and beginner tutorials
- [API Reference](https://docs.nvidia.com/nemo/microservices/latest/api/index.html.md): Comprehensive API documentation for all microservices
- [Tutorials](https://docs.nvidia.com/nemo/microservices/latest/get-started/tutorials/index.html.md): Step-by-step guides for various workflows

## Core Concepts

- [Platform Concepts](https://docs.nvidia.com/nemo/microservices/latest/about/core-concepts/platform.html.md): Foundational concepts of the NeMo platform
- [Customization Concepts](https://docs.nvidia.com/nemo/microservices/latest/about/core-concepts/customization.html.md): Fine-tuning concepts for model customization
- [Evaluation Concepts](https://docs.nvidia.com/nemo/microservices/latest/about/core-concepts/evaluation.html.md): Concepts for evaluating AI workflows
- [Inference Concepts](https://docs.nvidia.com/nemo/microservices/latest/about/core-concepts/inference.html.md): Concepts for testing and serving custom models
- [Guardrail Concepts](https://docs.nvidia.com/nemo/microservices/latest/about/core-concepts/guardrails.html.md): Concepts for controlling AI workflow interactions

## Optional

- [Release Notes](https://docs.nvidia.com/nemo/microservices/latest/about/release-notes/index.html.md): Latest updates and changes
- [EULA](https://docs.nvidia.com/nemo/microservices/latest/eula.html.md): End-user license agreement
- [Key Features](https://docs.nvidia.com/nemo/microservices/latest/about/key-features.html.md): Features mapped to AI model development workflow stages