All Projects
MLOps

Multi-Cloud AI Deployment

Python tool that deploys open-source models (Llama 3, BERT, GPT variants) across AWS, GCP, and Azure through one unified API.

Role

Backend Developer

Timeline

2024

Stack

PythonFastAPIDockerHugging FaceAWS SageMakerGoogle Cloud Vertex AIAzure MLWebSocketsAWS S3OpenAIModel DeploymentMLOpsContainerization

Overview

A Python-based AI model deployment tool that simplifies deploying open-source models across AWS SageMaker, Google Cloud Vertex AI, and Azure Machine Learning. The platform provides a unified API eliminating the complexity of cloud-specific documentation. Features include automated model lifecycle management, real-time deployment feedback via WebSockets, and support for popular models including Llama 3 and Hugging Face transformers.

Architecture & Decisions

Unified Multi-Cloud Model Deployment

Simplifies deployment of Hugging Face models including Llama 3, BERT, and GPT variants across three major cloud platforms with a single unified interface. Handles model artifact storage via AWS S3, automated Docker containerization, and cloud-specific configuration without requiring deep platform expertise. Supports model creation, querying, fine-tuning, and deletion operations.

Real-Time Deployment Monitoring

Features WebSocket-based real-time feedback for deployment progress and logs with informative status messages. FastAPI-powered backend ensures high-performance API routing for model lifecycle operations. Provides consistent deployment experience across AWS SageMaker, GCP Vertex AI, and Azure ML, dramatically reducing time to production for AI/ML teams.

The Hard Part

Every cloud ML platform has its own deployment quirks and documentation. The hard part was abstracting that into one interface without hiding the control teams actually need over instance type, scaling, or model config.

Outcomes

Interface

One unified API

Models

Llama 3 · BERT · GPT

  • WebSocket-based real-time deployment feedback and logs
  • FastAPI backend for high-performance model lifecycle routing
  • Automated Docker containerization and S3 artifact storage
  • Consistent deploy experience across SageMaker, Vertex AI, and Azure ML

Up Next

View all
SaaS

Travel Discount Card Ecosystem

Discount card ecosystem for a travel agency — admin panel, consumer app, and e-commerce integration in one platform.

Read case study

LET'S WORK TOGETHER

Have a system that needs to ship like this?

Get in touch