Madhur Prashant

Lead Generative AI/Agents Solutions Architect, AWS

Architecting the future of AI with cutting-edge generative models, intelligent agent systems, and enterprise-grade AI solutions on AWS. Transforming complex business challenges into innovative AI-powered solutions.

About Me

As a Lead Generative AI/Agents Solutions Architect at AWS, I specialize in designing and implementing sophisticated AI systems that push the boundaries of what's possible with artificial intelligence. My expertise spans across multi-agent systems, foundation model benchmarking, incident response automation, and enterprise-grade AI applications.

With a focus on practical, production-ready implementations, I help organizations leverage the full potential of AWS AI services including Amazon Bedrock, SageMaker, and advanced agent frameworks like LangGraph and Strands AI.

Generative AI

Foundation models, fine-tuning, and optimization

Multi-Agent Systems

Complex agent orchestration and collaboration

AWS AI Services

Bedrock, SageMaker, and cloud-native AI

AI Safety & Security

Guardrails, responsible AI implementation

Featured Projects

A showcase of cutting-edge AI and agent development projects

A2A Multi-Agents AgentCore

Sophisticated multi-agent system for automated AWS operations, monitoring, and incident management with agent-to-agent communication.

AWS Bedrock LangGraph Streamlit CloudWatch

Foundation Model Benchmarking Tool

Comprehensive benchmarking tool for Foundation Models across AWS platforms, comparing performance and determining optimal serving stacks.

SageMaker Bedrock TensorRT HuggingFace

AWS Support Agentic AI App

Secure telehealth generative AI assistant with advanced safety guardrails, demonstrating responsible AI implementation for healthcare.

Bedrock Guardrails Claude 3.5 FAISS Strands AI

Meta Tools and Agents

Meta-agentic system enabling AI agents to dynamically create and manage their own tools with autonomous development capabilities.

Strands SDK LangGraph Amazon Bedrock Claude AI

Strands Multi-Agentic System

Multi-agent AI system for comprehensive financial analysis using collaborative specialized agents in a hierarchical architecture.

Strands SDK AWS Bedrock Amazon Nova Jupyter

Build Agents with LangGraph

Comprehensive tutorial and implementation guide for building intelligent AI agents from simple systems to complex architectures.

LangGraph FastAPI AWS Lambda Streamlit

Letta Agent Samples

Demonstration of stateful AI agents with persistent memory that learn from interactions and manage context intelligently.

Python Docker AWS Bedrock Anthropic

AWS CloudGuard MCP

Intelligent AWS incident response system using Claude Model Context Protocol with automated monitoring and natural language interactions.

CloudWatch Claude AI Jira API Python 3.12

LangGraph AWS Incident Response

Autonomous AWS incident response system using multi-agent AI workflows with graph-based agent coordination.

LangGraph LangChain AWS SDK Jupyter

Unsloth Finetune EC2

Optimized language model fine-tuning with 2x speed improvement and 70% memory reduction using Unsloth framework on AWS GPU instances.

Unsloth HuggingFace AWS EC2 Jupyter

DeepSeek SGLang AWS

High-performance DeepSeek language model deployment using SGLang inference system with advanced runtime optimization techniques.

SGLang DeepSeek R1 Docker Python 3.12

Let's Connect

Interested in AI innovation, collaboration, or have questions about my projects?

Role

Lead Generative AI/Agents Solutions Architect at AWS