Welcome to the Scale’s GenAI Platform (SGP) documentation!
- In these Guides pages, read about our various Components and use our Tutorials to easily get started with our APIs.
- Refer directly to our API Reference page in our top nav bar for comprehensive documentation on how to use our REST API.
Overview
The Scale GenAI Platform empowers modern enterprises to rapidly develop, test and deploy Generative AI applications for custom use cases, using their proprietary data assets. It includes an API, SDK and web frontend which abstract the flexible use of both open and closed-source resources, providing full-stack capabilities that meet enterprise security and scalability standards.Your Data
Optimizing LLMs starts with your data. Connect popular data sources and transform your data with the Scale Data Engine to implement optimized RAG pipelines and models fine-tuned for your domain-specific tasks.Any Model
Customize, test, and deploy all major closed and open-source foundation, embedding, and reranking models from OpenAI, Cohere, Meta, and more.Your Cloud
Securely customize and deploy enterprise-grade Generative AI applications in your own VPC, including AWS and Azure. Now available on the Azure Marketplace.Key Features
Advanced Retrieval Augmented Generation (RAG)
- LLMs can accurately reference your knowledge base with Scale’s tools for optimized Retrieval Augmented Generation (RAG).
- Convert knowledge base data into embeddings to serve as long-term memory, which the model can retrieve.
- Our comprehensive toolset includes data connectors, custom embedding models, vector stores, chunk summarization, chunk and metadata extraction, advanced reranking, and RAG and reranker fine-tuning.
Build Custom Models
- Fine-tune LLMs using your proprietary data or Scale expert data to improve performance & reliability on your unique use cases, while reducing latency and token consumption.
- Choose from any leading closed or open-source foundation models, including OpenAI’s GPT-4o Cohere’s Claude 3.5, and Meta’s Llama 3.1 and more.
- Leverage the Scale Data Engine to transform your data, and generate the highest quality training data for any use case.
Testing and Evaluation
- Optimize the performance of your applications by testing different data, prompts, RAG pipelines, models, and fine-tuning strategies.
- Get immediate insights into the quality of your AI with out of the box Scale Report Card metrics, which reliably assess key areas like accuarcy, quality and trust & safety, building on Scale’s years of experience creating high quality AI systems.
- Compare and evaluate base models and customized completion, embedding, and reranking models to determine the best model mix for your use case.
- Perform automated and human-in-the-loop benchmarking of the performance, reliability, and safety of your customized models or entire Generative AI applications.
- Create and manage test cases, define evaluation metrics, perform evaluations with subject matter experts, and analyze evaluation results.
Deployment and Monitoring
- Trace all calls to SGP endpoints with detailed engagement and execution metrics
- Chart usage and performance over time with intuitive charts and drill-downs
- Run online evals using Scale standardized metrics like faithfulness, coherence and context recall.
Enterprise-Ready
- Securely customize and deploy Generative AI applications in your own VPC, including AWS and Azure.
- Enterprise-grade RBAC and SAML SSO built-in.
- Secure centralized management of API keys.
- To get started with this, please contact our sales team.