David Plappert
Principal AI & Cloud Architect · AWS Consultant
Agentic AI running in production on AWS
Available for consulting & contract engagements — production AI engineering on Amazon Bedrock, reliability and observability, AWS architecture review, and vendor platform oversight
since EC2 beta
cloud architecture
Consulting Services
Senior AWS depth, delivered as focused engagements. Every service below is something I do in production today — not a practice area on a slide.
AI Engineering on AWS
Agents and retrieval that run in production, not demos. Amazon Bedrock (Converse API, tool use), RAG over pgvector and OpenSearch Serverless, multimodal document extraction, and agentic CI/CD wired to your stack through custom MCP servers. I run all of it in production today.
Reliability & Observability Engineering
SLOs and error budgets, load testing, monitoring and alerting design, and high-availability and disaster-recovery architecture for revenue-bearing AWS systems — plus observability that stays affordable at hundreds of gigabytes of logs a day.
AWS Architecture Review & Well-Architected Assessment
Independent, senior-level review of your AWS architecture against the Well-Architected Framework — reliability, security, cost, and operational excellence — with a prioritized, actionable findings report.
Vendor Platform Oversight (Buy-Side Review)
Paying a vendor seven or eight figures for a platform? I am the sole client-side technical counterparty on exactly that arrangement today for a Fortune 50 manufacturer — reviewing every vendor-shipped release (CDK, CloudFormation, Terraform) before it reaches production, and turning findings into corrective action the vendor actually ships.
Cloud Migration Strategy
On-premises to AWS, monolith to microservices, or platform to platform. I led the fastest state-based health exchange transition off the federal FedHub on record, and Best Buy’s billion-dollar e-commerce systems onto AWS.
Fractional Principal Cloud Architect
Ongoing senior architecture leadership on a retainer basis — roadmap, design reviews, vendor management, and mentoring for teams that need principal-level AWS depth without a full-time hire.
Why teams hire me
I ship AI to production, not demos. For a Fortune 50 manufacturer I built an observability layer on Amazon Bedrock that finds production defects, opens a pull request against the CDK with a proposed fix, and pages the engineer who owns it — Bedrock Converse API for tool calling, with embeddings in pgvector for semantic search across logs, runbooks, and architecture docs. Alongside it runs a Claude Code workflow that handles deploys, dry runs, bug detection, and CI/CD, with several agents in parallel bridged to GitHub and AWS through custom MCP servers I wrote.
I own reliability where downtime is measured in dollars per minute. Today that is a real-time 3D commerce platform earning ~$3,000 a minute when it is running: I designed its load testing, SLOs, monitoring, and alerting, scaled it from 2,000 to 8,000 concurrent users, and keep 330+ GB of logs a day both useful and affordable. I am also the sole client-side technical counterparty on the eight-figure vendor contract behind it. The same through-line runs back to Best Buy's e-commerce infrastructure on Thanksgiving Day 2012 — 25% of all US e-commerce traffic, one engineer on site: me.
The depth goes back 19 years, since the EC2 beta. I am the sole inventor of US Patent 11,146,689, an event-driven serverless architecture on AWS, and the DevOps team I led won the AWS City on a Cloud Award twice (2016, 2018) for the DC Health Benefit Exchange architecture I owned. I spent a decade as a client-facing principal architect at a government health-tech consultancy, where architecture I scoped and delivered grew engagements by $1M and $2M per year, payment workflows I designed processed $500M+ annually under HIPAA, SOC 2, and IRS Publication 1075, and I advised 18F (GSA) on state health-exchange replication.
Based in Peoria, Illinois · Remote, US-wide · Schedule a call