For 16 years I've designed the cloud platforms, AI agents, and code that enterprise teams put into production. Hands-on across architecture, AI engineering, and full-stack delivery — at the scale where it actually matters.
Most "cloud transformations" are PowerPoint deep. Most AI demos die before they reach production. The work I do is the other kind.
Across 16 years I've embedded with enterprise teams — FIS, JPMorgan Chase, Microsoft, HSBC, Banco Itaú — turning ambiguous problems into shipped systems. Recently that has meant designing an AWS Data Mesh serving 9M+ wealth management clients, building production AI agents on Amazon Bedrock, and leading large-scale migrations from on-prem to cloud-native.
What I bring that most architects at this level don't: I'm still hands-on. I read the code. I write the migration scripts. I'm an active full-stack engineer in Java and Python, not a PowerPoint architect. Currently going deeper on Agentic Architecture and preparing for the Claude Architect certification.
Cloud, AI, data, and code — the four threads that run through every project I take on. Tools change. The thinking doesn't.
AWS-first, Azure-fluent. Well-Architected Framework reviews, multi-account governance, large-scale migrations, FinOps.
Production AI agents on Bedrock and GPT-4. RAG pipelines, knowledge bases, guardrails, evals, and observability.
Event-driven architectures, Data Mesh design, streaming pipelines, and governance at enterprise scale.
Hands-on, not just hands-off. I still read code, write code, and ship code — in Java, Python, and the frameworks they need to live in.
Open source implementations of patterns from my production work. Each repo solves a real architecture problem and includes the decisions behind it.
An agentic AI security operations system on AWS. Built with Bedrock, Lambda, and a multi-step reasoning pipeline. Reference implementation for putting AI agents into production with guardrails and observability.
Terraform-based reference architecture for a governed Data Mesh on AWS — domain-driven ownership, federated governance, and event-driven product integration. Inspired by the platform I built for 9M+ wealth management clients.
A production-pattern Spring Boot service demonstrating clean architecture, API gateway integration, and the operational scaffolding (observability, security, CI/CD) that real microservices need.
Hands-on architecture labs covering Well-Architected pillars, event-driven systems, IaC patterns, and AI engineering experiments.
I write about the decisions behind production systems — what works, what breaks, and what I'd do differently. Published on Medium.
Most systems don't struggle because of poor technology choices. They struggle because responsibilities don't scale as the system grows.
From events to data products — how to design a Data Mesh that actually delivers domain-driven ownership in production.
A practical guide to picking the right agent framework when you actually have to put it into production.
Most AI architecture diagrams follow the same toy pattern. Here's what changes when you have to actually run it in production.
Applying the AWS Well-Architected Cost Optimization pillar to a secure GraphQL data platform — cost as an architectural intent, not an afterthought.
Determining optimal Lambda memory configuration as a deliberate design choice, not a guess.