Agentic AI Architect · Author · Builder

Abhijeet Patil

Senior Solutions Architect at Amazon Web Services building agentic AI systems for enterprise customers. Author of the OpenAPI MCP Server in awslabs/mcp, co-author of the Apress book Practical Agentic Coding, and co-founder of AgenticCoders.dev. 20+ years architecting distributed platforms and agentic AI systems for Fortune 50 enterprises, backed by 11 AWS certifications.

LinkedIn GitHub Medium Substack Book ResearchGate Email

About

Executive Summary

I design and build agentic AI systems for enterprise customers, from MCP infrastructure in awslabs to Bedrock AgentCore reference implementations to industry-specific LLM frameworks. As a founding program member of AWS's CDE (Customer Development Engineering) program, I delivered two business-critical agentic AI prototypes for enterprise customers.

I design and rescue internet-scale distributed systems: multi-region platforms serving 100M+ users and billions of events, architected to 99.99% availability targets, across Financial Services, Insurance, and Hospitality. My work tends to start where other teams have run out of options: the chronic failure nobody could reproduce, the latency nobody could explain. Backed by 11 AWS certifications, including Solutions Architect Professional, DevOps Engineer Professional, Machine Learning Specialty, and Security Specialty.

Beyond my day role, I'm the co-author of the Apress book Practical Agentic Coding, co-founder of AgenticCoders.dev, and author of DevMemory, a published memory architecture for AI coding agents. I serve on the AI Governance Committee of the AI Hospitality Alliance.

I'm completing an Executive MBA with a Cybersecurity focus at UNC Wilmington in Fall 2026, and I'm an active contributor to the Generative AI and Machine Learning Technical Field Community at AWS.

I mentor architects and engineers, and deliver AI-DLC workshops helping enterprise teams accelerate agentic AI delivery.

Expertise

Areas of Focus

Internet-Scale Architecture

AWS, Azure, IBM Cloud, Hybrid/Multi-Cloud strategies, migration, and modernization at enterprise scale.

Agentic AI Engineering

MCP servers, Bedrock AgentCore, multi-agent orchestration, agent memory, tool calling, agent identity, and domain-specific LLM frameworks.

Cybersecurity

Security architecture, compliance frameworks, AWS Security Specialty certified, EMBA Cybersecurity focus.

AI Governance & Validation

3-layer AI defense for production LLMs. Cedar Policy Engine for sub-millisecond deterministic validation, Bedrock Guardrails, AI-agent permission frameworks, and enterprise trust controls. Read the framework →

Executive Influence

Trusted advisor to CTO and VP/SVP leadership. Framing technical investment as business capability, defining multi-year technical vision, and building the architects who carry it.

Performance & Observability

Full-stack observability, OpenTelemetry (ADOT), distributed tracing, APM telemetry, and latency optimization at scale.

Technology

Technical Proficiency

Cloud Platforms

AWS Microsoft Azure IBM Cloud Red Hat OpenShift Kubernetes

Agentic AI

MCP Bedrock AgentCore Multi-Agent Orchestration Agent Memory Agent Identity / OAuth Agentic Coding (Claude Code, Kiro, Antigravity, Copilot) AI-DLC

AI / Machine Learning

Amazon Bedrock SageMaker Domain-Specific LLMs Fine-Tuning RAG LangChain Vector Databases

Languages & Frameworks

Python Java Spring Cloud Node.js TypeScript Terraform AWS CDK

Infrastructure & DevOps

Microservices Docker Event-Driven Architecture OpenTelemetry CI/CD HA/DR

Career

Professional Experience

Client names are abstracted where the engagement is not public. Named references and a detailed CV are available on request.

2022 to Present

Senior Solutions Architect

Amazon Web Services (AWS) · Raleigh, NC

  • Lead AWS teams helping enterprise customers build agentic AI platforms and AI capabilities on Amazon Bedrock AgentCore and AWS AI services, and deliver AI-DLC workshops that accelerate their teams from prototype to production
  • Operate as a principal-level technical authority advising CTOs and VP Engineering at Fortune 50 enterprises on AI adoption strategy, cloud-native architecture, and scalable ML inference pipelines
  • Led zero-to-one architectural rebuild for a top-tier enterprise, delivering real-time booking and profile management via Model Context Protocol (MCP) for household AI devices
  • Engineered a multi-industry AI platform on AWS Bedrock AgentCore and Strands SDK, adopted as the canonical reference architecture used by AWS field teams across the Travel & Hospitality vertical
  • Called in to unblock a stalled enterprise identity rollout under a hard deadline: designed and proved a federation architecture in days, then handed the pattern to the customer's platform team
  • Authored the OpenAPI MCP Server in awslabs/mcp (9,400+ stars), turning any OpenAPI-described enterprise API into MCP tools for AI agents; own its ongoing releases including security hardening (SSRF protection, credential isolation) and FastMCP 3.x migration
  • Led the HospitalityLLM industry model framework within the AWS Travel & Hospitality Enterprise Architecture team, a domain-specific LLM framework for the hospitality industry, published on the AWS Builder Community and by the AI Hospitality Alliance
  • Built and published customer-facing agentic AI reference implementations on aws-samples: Bedrock AgentCore Gateway user federation (OAuth 2.0 APIs as MCP tools) and an AI-powered observability agent for incident investigation
  • Coach and mentor fellow architects; active in the Generative AI and Machine Learning Technical Field Community
2016 to 2022

Executive IT Architect / Hybrid Cloud Leader

IBM Corporation · Research Triangle Park, NC & Boston, MA

  • Designed and directed a practitioner retraining initiative certifying 300+ engineers in Full Stack and Cloud Architecture
  • Led cross-functional task force for a highly escalated global trade finance platform, cutting p99 latency by 93% (a ~14x reduction in tail response time), measured from deep APM telemetry, recovering a delivery escalated to IBM GBS North America leadership
  • Architected the largest Digital Platform modernization for a global insurance leader, multi-region Microservices SaaS across 40+ countries, increasing deployment speed-to-market by 60%
  • Chief Solution Architect (CIO Office) for Online, Mobile, and ATM channels at a major US banking enterprise
2006 to 2016

Lead Developer to Group Manager

IBM India Pvt. Ltd. · Pune, Mumbai, Chennai (India) & Charlotte, NC & Delaware & Colorado

  • Definitive technical authority for a major e-commerce platform (30M+ active digital users); resolved a chronic memory leak causing dozens of cascading server failures monthly, isolating it to the platform's core framework via single-threaded heap analysis and rewriting the flaw, after client engineering teams had exhausted all options
  • Scaled offshore engineering teams from zero to full production with rigorous CI/CD, DevOps, and cloud-native competency pools
  • Core backend workflow platforms (SOA/REST) handling millions of transactions for global financial institutions

Publications

Book & Research

Practical Agentic Coding Book Cover
Upcoming Book · Apress (Springer Nature)

Practical Agentic Coding

Beyond Vibe Coding and AI Assistants for Successful Autonomous Software Development

Foreword by Dr. Matt Wood, Chief AI and Technology Officer, AWS (formerly CTIO, PwC)

Co-authored guide on building production-grade autonomous coding systems, covering agentic AI architectures, evaluation frameworks, and enterprise deployment strategies.

Available to Pre-Order in US & Europe

US
Research Preprint · Zenodo · DOI 10.5281/zenodo.19363278

DevMemory: A Structured Memory Management Architecture for Persistent Knowledge in AI Coding Agents

A memory architecture for AI coding agents that classifies knowledge into episodic, semantic, and procedural types across agent, project, and organizational scopes, with a formal trust-scoring model (provenance, recency, contradiction, and reinforcement signals). Implemented as an open-source MCP server with 325 automated tests; benchmarked at 87.5% cross-session information reuse and 27-91% token savings under budget constraints.

Read on Zenodo
Governance Architecture Overview: Autonomous Development
Research Paper · ResearchGate

Governing Agentic AI: A Strategic Framework for Autonomous Systems

A strategic framework addressing governance, accountability, and oversight challenges in the deployment of autonomous AI systems at enterprise scale.

Read on ResearchGate
AI Insight Paper · AI Hospitality Alliance · Jul 2026

Why Travel and Tourism Needs More Than a Generic LLM

Industry insight paper introducing a five-part domain-fitness framework for deciding when hospitality businesses need domain-specific fine-tuned models over generic LLMs with retrieval, and proposing a multi-agent architecture (revenue management, guest experience, operations) coordinated by an orchestrator. I also serve on the Alliance's AI Governance Committee.

Read on AI Hospitality Alliance

Thought Leadership

Writing & Ideas

Published perspectives on agentic AI, domain-specific LLMs, cloud architecture, and the future of technology.

Featured · Co-Founder & Podcast Creator

AgenticCoders.dev: Enterprise AI Knowledge Platform

Enterprise AI knowledge platform for practitioners, architects, and technology leaders building real AI systems. I co-founded it and produce the "Building Enterprise AI" series and podcast (Spotify, Apple Podcasts, YouTube), an AI-agent-hosted show covering domain-specific LLMs, agentic architectures, and production AI deployment.

Visit AgenticCoders.dev
Featured · Official AWS Blog

Choice Hotels Adopts Amazon Managed Service for Prometheus for Operational Excellence and Cost Efficiency

Published on aws.amazon.com. A deep dive into how Choice Hotels implemented Amazon Managed Service for Prometheus with AWS Distro for OpenTelemetry to achieve operational excellence and cost optimization in their monitoring infrastructure.

Read on AWS Blog
AgenticCoders · Building Enterprise AI Series

Why Your Enterprise Needs Its Own LLM (And Why It Costs Less Than You Think)

Part 1 of the "Building Enterprise AI" series: the case for domain-specific models over broad foundation models, grounded in a hands-on fine-tuning project and the real cost economics.

Read on AgenticCoders → Substack · Agent Observability

How We Learned AgentCore Observability the Hard Way

A human architect and an AI coding agent debug distributed tracing across 3 runtimes, 59 spans, and a silent 401 cascade.

Read on Substack → Agentic AI

Agentic AI: A Vehicle, Not the Destination

Why businesses should treat agentic AI as a means to solve real customer problems rather than a goal in itself, and why managing AI agents is closer to coaching than coding.

Read on Medium → AI / Society

The Last Generation That Questions AI Is Raising the First That Never Will

A look at how the next generation's relationship with AI will fundamentally differ from ours.

Read on Medium → AI / Education

How Today's Students Are Outpacing Industry Leaders in AI

Observations on the widening gap between how students and industry leaders adopt AI tools.

Read on Medium →

AWS Builder Community

AWS Builder · Industry LLM

HospitalityLLM: Building an Industry Model Framework

A framework for building domain-specific LLMs for the hospitality industry, developed with the AWS Travel & Hospitality Enterprise Architecture team.

Read on AWS Builder → AWS Builder · Identity

Filtering AWS Identity Store Users by Status: A Complete Guide

A hands-on guide to filtering and managing user status in AWS Identity Store for access governance.

Read on AWS Builder → AWS Builder · Operations

Automating AWS Systems Manager Parameter Store Backup and Cross-Region Replication at Scale

Automating parameter backup and cross-region replication for operational resilience and disaster recovery.

Read on AWS Builder → AWS Builder · AI Productivity

Boosting AWS Dev with Amazon Q CLI: My AI Coding Journey

Exploring how Amazon Q CLI accelerates cloud development workflows through AI assisted coding.

Read on AWS Builder → AWS Builder · Generative AI

Seamless AI Image Generation with Nova Canvas and MCP

Integrating Amazon Nova Canvas with Model Context Protocol for seamless AI image generation workflows.

Read on AWS Builder →
View all articles on Medium Substack Newsletter AWS Builder Community Profile

Agentic AI Systems

Built & Shipped

Original open-source work and customer-facing reference implementations in agentic AI: MCP infrastructure, agent memory, and production agent patterns.

Python · awslabs/mcp · 9.4k★ monorepo

OpenAPI MCP Server (Author)

Created and maintain the server that turns any OpenAPI-described API into MCP tools for AI agents. Own releases through v1.1.0: SSRF/LFI protection, credential isolation, FastMCP 3.x migration, and a CVE-free container image.
Python · MCP Server + Research

DevMemory (Author)

Persistent memory architecture for AI coding agents: episodic/semantic/procedural knowledge types, trust-scored retrieval, and token-budget packing. Open-source MCP server with 325 automated tests, published on Zenodo.
Industry LLM Framework · AWS

HospitalityLLM & LLM Command Center (Author)

Led an industry model framework for hospitality within the AWS Travel & Hospitality Enterprise Architecture team, and sole author of the LLM Command Center for running the full domain-LLM lifecycle in one place.

Article → Video walkthrough →
Python · aws-samples

Bedrock AgentCore Gateway User Federation (Author)

Built this reference implementation exposing OAuth 2.0 secured APIs as MCP tools for AI agents while preserving user authentication context, using AgentCore Gateway Interceptors: agent identity and credential propagation in practice.
Python · aws-samples

Observability Agent for Bedrock AgentCore (Author)

AI-powered observability agent that investigates production incidents alongside SREs to reduce mean time to resolution, applying agentic tool use to operational telemetry.
TypeScript

Hotel Revenue Optimization Agent (Author)

End-to-end revenue optimization agent for hospitality: LLM agent core with function-calling tools, configurable agent persona, and a React frontend, deployable on AWS (S3 + CloudFront, ECS/EKS), with automated UI and API test suites.

View all repositories

More on GitHub

Education

Academic Background

2024 to Fall 2026

Executive MBA

Cybersecurity Focus

Cameron School of Business, UNC Wilmington

2006

PG Diploma

Advanced Computing

C-DAC, Pune, India

2003

Bachelor of Engineering

Mechanical Engineering

Shivaji University, India

Credentials

Certifications & Recognitions

AWS Certifications

Professional

Solutions Architect

Professional

DevOps Engineer

Specialty

Machine Learning

Specialty

Security

Associate

Solutions Architect

Associate

Developer

Associate

SysOps Administrator

Associate

ML Engineer

Associate

Data Engineer

Foundational

AI Practitioner

Foundational

Cloud Practitioner

Other Credentials

AWS Badge

Agentic AI Demonstrated

Microsoft

Azure Architect (AZ-303)

Oracle

OCI Architect Professional

Red Hat

OpenShift Admin

Red Hat

OpenShift App Dev

Red Hat

API Management

IBM

Enterprise Design Thinking

Speaking & Mentorship

AWS

Sr. Certified Speaker

Represents the organization in external technical forums

IBM

Credly-Certified Mentor

Formally accredited mentorship of practitioners and architects

IBM

Recognized Educator

Designed the accreditation certifying 300+ practitioners

Creator and producer of the AgenticCoders podcast, an AI-agent-hosted show on YouTube, Spotify, and Apple Podcasts.

Topics I speak on: Agentic AI · AWS AI · Enterprise LLM · Agentic Coding · Developer Productivity. Invite me to speak →

Verified Credentials

All certifications independently verified on Credly

Verify on Credly

Connect

Let's Work Together

Whether you're building agentic AI systems, modernizing on the cloud, or exploring AI/ML adoption, I'd be glad to connect.

LinkedIn Email AgenticCoders GitHub Medium Substack