Code companion for Beyond a Single Agent: Subagent Delegation with GPT-6.1 Sol. The demo compares the same model working as a single agent and with subagents available through the Responses API’s b…

Mars mission control simulator built with the OpenAI Agents API, React, and TypeScript. Five AI agents coordinate mission decisions with human approval, supported by a Promptfoo evaluation harness …

Promptfoo evaluations of OpenAI reasoning quality, tokens, latency, and cost

This project compares reasoning-token usage, latency, and estimated on-demand cost across Anthropic Claude and OpenAI GPT models on Amazon Bedrock. Both benchmark scripts use the same six scenarios.

A question-answering agent built with the OpenAI Agents SDK against Amazon Bedrock Mantle's OpenAI-compatible endpoint, developed incrementally across six versions.

This project trains an XGBoost regression model with SageMaker Pipelines, registers the model in SageMaker Model Registry, and deploys the approved model to Amazon EKS through EventBridge and CodeB…

An interactive multi-agent mission simulator built with the OpenAI Agents SDK

Building a Claude Desktop plugin to access an Amazon Bedrock AgentCore Gateway for Web Search

AgentCore Web Search Gateway Demonstration

A Gradio web application for AI image generation and editing, powered by 4-bit quantized Qwen Image models.

Qwen3 TTS Demo

A full-stack multi-agent orchestration system demonstrating construction project management using AI agents.

Deploying Qwen Image Edit Model to Amazon EKS with GPU Acceleration

Intelligent image cropping tool with multiple detection methods including You Only Look Once (YOLO), DEtection TRansformer (DETR), Real-Time DEtection TRansformer (RT-DETR), Roboflow DETR (RF-DETR)…

Image editing using the Qwen Vision Language Model, deployed on Amazon EKS with GPU acceleration.

Deploy and performance test the state-of-the-art RF-DETR object detection model to Amazon SageMaker AI using PyTorch.

Fine-Tuning Wav2Vec2 for Real-Time Deepfake Audio Detection: From custom dataset preparation and Wav2Vec2 fine‑tuning to SageMaker deployment and streaming inference

Creates and visualizes 200 text embeddings at 512 dimensions each, projected into 2D for visualization, across four popular open weight text embedding models

Learn how to deploy and performance test the state-of-the-art YOLO11 object detection model to Amazon SageMaker AI for production-ready, real-time inference with GPU acceleration.

Building an AI-Powered Smart Image Cropping Pipeline with YOLO, DETR, and RT-DETR

Create short product videos from keyframes using Amazon Bedrock, Luma AI, CloudScape, and AWS Fargate

Personalized Recommendation Agent from demonstration at AWS LA Summit 2025

Demonstrating the use of Amazon Nova Multimodal Embeddings and TwelveLabs Pegasus 1.2 models on Amazon Bedrock along with Amazon OpenSearch Serverless to perform semantic search.

Building Smarter Content Recommendations with PCA, K-Means, and Generative AI, on Ollama and Amazon Bedrock

Build a free web-based research agent using Strands Agents, Ollama running Qwen3, and the Tavily MCP Server

Demonstration of Strands Agents tool calling capabilities with Ollama and IBM Granite 4.0 SML.

Demonstration of a Video Search Agent build with AWS Strands Agents, TwelveLabs on Amazon Bedrock, OpenSearch, and Gradio.

How TwelveLabs AI Models on Amazon Bedrock and OpenSearch enable multi-vector semantic search for video content.