agent-platform-tuning

What problem does it solve?

This Skill eliminates the complexity of manually managing the end-to-end workflow for fine-tuning large language models (both open-source and Gemini models) using Agent Platform infrastructure, which traditionally requires handling environment setup, dataset preparation, job configuration, monitoring, and deployment across multiple tools and services.

Core Features & Use Cases

• End-to-End Workflow Guidance: Step-by-step instructions for every phase of model tuning, from initial environment and IAM setup to final model deployment, with built-in validation checks at each stage.
• Dual Model Support: Handles tuning workflows for both open-source models (Llama, Gemma, Qwen) and Gemini models, with model-specific recommendations for hyperparameters and hardware.
• Use Case: A machine learning engineer can use this Skill to fine-tune a Llama 3.1 8B Instruct model on their internal customer support dataset, validate the dataset format, estimate tuning costs, submit the job, monitor progress, and deploy the tuned model to production without switching between multiple documentation sources.

Quick Start

Use the agent-platform-tuning skill to fine-tune your custom domain-specific LLM on a prepared dataset and deploy the tuned model via Agent Platform.