Sagify: Machine Learning Workflow Management

Key Features

• Simplified SageMaker Interface

  Provides CLI commands to support resource provisioning, distributed training, and cloud deployment on AWS SageMaker.

• LLM Gateway API

  A FastAPI-based interface that supports interactions with proprietary and open-source large language models.

• Hyperparameter Optimization

  Supports Bayesian Hyperparameter Optimization on AWS SageMaker using JSON configuration files.

• Batch Inference

  Supports processing of large data volumes using JSONL files stored in S3 buckets for offline embeddings and predictions.

• Lightning Deployment

  Supports deploying pre-trained models from frameworks such as sklearn, HuggingFace, and XGBoost.

Use Cases

• Custom ML Model Training

  Training a model locally and pushing it to AWS ECR for cloud training on SageMaker.

• Unified LLM Integration

  Using a single API to access proprietary LLMs (OpenAI, Anthropic) and open-source models.

• Large-scale Embedding Generation

  Running batch inference jobs to generate embeddings for use in vector databases or search systems.

• Model Deployment

  Deploying foundation models or custom-trained models as RESTful inference endpoints.

Integrations

• AWS SageMaker
• AWS ECR
• AWS S3
• OpenAI
• Anthropic
• FastAPI

FAQ

What is Sagify used for?

Sagify is used to manage machine learning workflows on AWS SageMaker, helping engineers train, tune, and deploy models while simplifying infrastructure management.

Does Sagify support non-AWS LLMs?

Yes, its LLM Gateway provides a unified interface to interact with proprietary models from providers like OpenAI and Anthropic, as well as open-source models on SageMaker.

What are the technical requirements for Sagify?

Users require Python (versions 3.7 through 3.11), Docker, and a configured AWS CLI.