I build the serverless platforms and automation an operations organization runs on.

Full stack engineer on AWS: event-driven services, compliance automation, and AI tooling built on MCP, Bedrock and LangGraph.

Email me Selected work

Continuous adaptation, refinement, and improvement.

More than 60 services and automations shipped to production, several of them running continuously for five years or more.

I started on Windows servers and kept evolving. Systems administration became SharePoint engineering. SharePoint engineering became leading a team of engineers and consultants supporting hundreds of customers. Then I transitioned back to delivering custom solutions.

Since then, my focus has been on building and delivering reliable, scalable, event-driven, decoupled, serverless applications — API Gateway, Lambda, Step Functions, DynamoDB, SQS, ECS, ECR, Fargate, OpenSearch — provisioned entirely through Terraform and deployed by automated pipelines.

Many of these solutions are required to be PCI DSS compliant. I have multiple years of experience implementing requirements, and creating and presenting evidence to auditors.

Most recently I've been working in the AI space — building Model Context Protocol servers, and delivering full stack solutions for clients on top of AWS Bedrock and Palantir.

Things I get called in for

AI tooling & MCP
Model Context Protocol servers that expose internal APIs to LLM assistants with server-side access controls — backed by Amazon Bedrock, and paired with agentic solutions built on LangGraph. Custom solutions built on top of Palantir. AWS Certified AI Practitioner. Palantir Data Engineer Certified.
Serverless architecture
Event-driven services built from API Gateway, Lambda, Step Functions, DynamoDB, SQS and SNS. Designed for fault tolerance first: idempotent handlers, dead-letter queues, and failure alerting on every state machine.
Platform APIs & front ends
Versioned REST APIs with custom Lambda authorizers, OpenAPI documentation and Athena-queryable request logs — paired with React and TypeScript single-page front ends that engineers actually use during an incident.
Data & analytics pipelines
Ingestion and transform pipelines feeding OpenSearch, Kinesis streams and Palantir Foundry — ending in dashboards used for leadership and business-review reporting. Palantir certified Foundry Data Engineer.
Infrastructure as code
Terraform or CloudFormation from the VPC up. Reusable modules, plan-and-approve pipelines with manual gates for production, and disposable per-branch test environments that tear themselves down.
Compliance engineering
PCI DSS subject-matter expert. I build the controls rather than document them after the fact: continuous monitoring against policy defined as code, with reviewable evidence generated automatically.

Systems I designed, built and operated

Nearly all of my professional code lives in private repositories, so what follows describes each system by its problem, approach and outcome. Happy to go deeper on any of these in conversation.

AI & MCP backed troubleshooting and analysis tool

Stack

  • AWS Bedrock
  • LangGraph
  • Python
  • DynamoDB
  • OpenSearch
  • API Gateway
  • React
  • Terraform
  • Vite

Tickets that pass between several teams get long, and engineers were spending too much time reading them before they could start diagnosing. This assistant works alongside the ticketing workflow: it gathers relevant context from infrastructure, knowledge bases and supporting APIs, and uses Amazon Bedrock with custom MCP servers to produce a compact summary, suggested troubleshooting steps with a confidence score, and draft remediation where appropriate. The summary regenerates as the ticket changes, so it is never stale.

  • Shortens time to diagnosis and helps engineers at every experience level work issues independently.
  • A LangGraph pipeline runs the gathered context through role-specific prompts for operations and software engineering perspectives.
  • MCP servers and search indexes supply the model's context; a caching layer keeps repeat analysis cheap.
  • A React front end presents actionable next steps rather than raw analysis.

MCP servers for LLM-assisted cloud operations

Stack

  • Python
  • Model Context Protocol
  • AWS Bedrock
  • LangGraph
  • FastAPI
  • Poetry
  • Terraform
  • LLM tooling

Giving an AI assistant useful access to cloud tooling is mostly an authorization problem, not a prompting one. I built an internal tooling API and the Model Context Protocol server that fronts it, backed by Amazon Bedrock, so engineers can investigate issues with an assistant that pulls in relevant context and suggests next steps, shortening investigations. A second server connects operational data to a LangGraph workflow that removes most of the manual preparation before customer business reviews, including feedback from simulated executive and architect perspectives. The same server also powers the troubleshooting assistant described above.

  • Access controls enforced on the server rather than left to the model.
  • A deliberately narrow tool surface — capability granted by design, not by prompt instruction.
  • Deployed as infrastructure as code alongside the API it wraps.

More systems

  • Automated alarm triage & remediation pipeline Repetitive first-response diagnostics now run on their own.

    Engineers were repeating the same diagnostic steps for recurring monitoring alarms. This service consumes the monitoring event stream, filters out noise, and routes qualifying alarms to Step Functions workflows that run diagnostics and, where it is safe to, remediation. The engineer picking up the issue gets a summary of what was found and what was done.

    • Handlers are organized per alarm type, so new conditions can be added without touching the pipeline.
    • Filtering early keeps cost proportional to actionable events rather than total event volume.
    • Every workflow branch reports its outcome, failures included, so nothing is silently dropped.
    • Extended beyond AWS to other cloud and hosted environments on the same workflow model, taking it company-wide.
    • Saves a substantial amount of engineering time every year.

    Stack Python, AWS Step Functions, Lambda, DynamoDB, AWS SQS, Terraform

  • Multi-cloud support tooling automation platform Standard tooling deployed across a multi-cloud fleet through one audited API.

    A platform for deploying standardized configuration and monitoring tooling across environments on more than one cloud provider. A catalog defines what can be deployed; a versioned REST API and worker fleet handle deployment, updates and removal; every request is logged for audit. New environments get a consistent baseline automatically, so engineers can start work without manual setup.

    • Serverless API with published OpenAPI documentation.
    • Structured, queryable request logs, so auditing a change is a simple SQL query.
    • React front end for self-service use.
    • A large, versioned catalog of supported automations.

    Stack Python, API Gateway, Lambda, DynamoDB, Athena, React, TypeScript, Vite, Terraform

  • Maintenance window API and front end Planned maintenance no longer generates noise for the engineers on call.

    Planned maintenance was generating alerts and tickets for conditions everyone already expected. This product lets teams schedule maintenance windows through an API and a React front end, replacing a workflow that previously meant scripting against an API by hand. Expected noise stays out of support queues, so engineers spend their time on real issues.

    • Serverless API with a React and TypeScript front end.
    • Coverage is automatic rather than opt-in, so it never depends on someone remembering to enroll.
    • Rebuilt on React and Vite, which made it significantly faster to load, and deployed through Terraform and GitHub Actions.
    • Delivered as a micro-frontend alongside existing tooling.

    Stack Python, API Gateway, Lambda, DynamoDB, React, TypeScript, Vite, Terraform

  • Continuous compliance monitoring with audit trail Privileged activity in regulated environments is reviewed continuously, not reconstructed after the fact.

    Regulated environments have to show that privileged activity was expected and reviewed. This service evaluates infrastructure activity against a policy defined as code, alerts on anything outside it, and produces a concise review of the exceptions, so the team meets PCI DSS log-review requirements without manually combing through full logs.

    • Policy is declarative and reviewed like any other code change.
    • Automated infrastructure changes are held to the same policy as human activity.
    • Built so that processing errors surface for review rather than disappearing.
    • Gives assessors a timestamped, reviewable trail.

    Stack Python, CloudTrail, DynamoDB, Lambda, PCI DSS, Terraform

  • Multi-cloud account assessment & reporting tool A multi-day manual cloud assessment became an automated report.

    Answering "what is actually running in this environment, and is it configured the way we recommend?" was a multi-day manual exercise. This tool automated it: requests are expanded into collection jobs, inventory and configuration are gathered across accounts, results are processed asynchronously with progress tracking, and the output is rendered as a report. Built first for AWS, then extended to Google Cloud.

    • Asynchronous job model with progress reporting, so long-running scans don't hold a request open.
    • Pluggable output writers for customer-facing and internal formats.
    • Parallel AWS and GCP implementations sharing the job and reporting model.

    Stack Python, AWS API Gateway, Lambda, AWS S3, AWS SQS, Step Functions, Terraform, Vite

  • Executive analytics pipelines and dashboards Scattered spreadsheets and reports consolidated into the single view leadership relies on.

    Data engineering and dashboards for leadership reporting on the Palantir Foundry platform: Python transforms that shape operational and business data, TypeScript server-side functions that compute the metrics, and React front ends built with Vite and Recharts for trend and business-review reporting. Also built the streaming transforms that bring automation events into the platform.

    • KPIs feed metric-specific LLM prompts that generate plain-language interpretations alongside the charts.
    • Transform layer separates raw ingestion from the metric definitions consumed by dashboards.
    • Server-side TypeScript functions keep metric logic in one place across dashboards and applications.
    • Front ends built against a shared internal design system for consistency with other tools.

    Stack Python, TypeScript, React, Vite, Recharts, Palantir Foundry, Kinesis

Work that's public

Most of what I write is closed source, but I have contributed to a number of open source projects.

  • cloudtools/troposphere

    Python library · Generates CloudFormation templates

    Contributor to the widely used Python library for generating CloudFormation templates. My merged work added support for AWS Auto Scaling Plans and Aurora backtrack windows, corrected the EC2 Spot Fleet load-balancer structure and an over-strict launch-template requirement, filled in missing CloudFront parameters and validators, and fixed a DynamoDB validator error.

  • Rackspace public Terraform modules

    Terraform modules · Reusable AWS infrastructure

    Author and long-term maintainer of several modules in Rackspace's public Terraform library. I wrote the SQS module outright, brought S3, SNS and SQS to parity with their CloudFormation equivalents, added Route 53 support and queue policies, fixed redrive-policy handling, reworked bucket ACL behavior to match AWS defaults after the provider's ACL changes, and set up their CI.

  • darktable

    Photo editing software · Camera support data

    Contributed Canon EOS R3 support to darktable, the open-source photography workflow and raw development application. Camera support there depends on contributors doing the measurement work: I shot and processed calibration reference frames to mathematically derive values for white-balance presets across the camera's color-temperature and preset range, plus a denoise profile covering its ISO range. Raised the tracking issue, then delivered both data sets.

  • Professional SharePoint 2013 Administration

    Contributing author · Wrox / Wiley · ISBN 978-1-118-49581-0

    Contributing author on Wrox's professional administration title. Alongside it, I presented pro bono at SharePoint Saturday community events, wrote and taught internal Rackspace University classes, and delivered customer-facing webinars.

  • AWS case study

    Solution highlighted · Amazon Web Services

    One of my internal services is described in the solution section of this AWS case study.

Certifications

Amazon Web Services

  • DevOps EngineerProfessional · 2024
  • Solutions ArchitectAssociate · 2016
  • DeveloperAssociate · 2021
  • SysOps AdministratorAssociate · 2024
  • AI PractitionerFoundational · 2025

Data platform & AI

  • Palantir Foundry Data Engineer2026
  • Palantir Foundry & AIP Builder Foundations2026
  • Palantir Foundry Aware2026
  • Generative AI with Large Language ModelsDeepLearning.AI & AWS · 2023

Microsoft

  • MCTS: SharePoint 2010
  • MCITP: Server AdministratorWindows Server 2008
  • MCSA: Windows Server 2008
  • MCTS: Active DirectoryWindows Server 2008
  • MCTS: Server Network Infrastructure
  • Microsoft Certified Professional

Other

  • FCC Amateur Radio Technician
  • Wilderness First ResponderNOLS Wilderness Medicine

Get in touch

Email is the reliable way to reach me, and I answer it. I'm happy to talk through any of the work above in more detail, and open to conversations about interesting problems in cloud platform engineering. Generally available for calls 2:00–7:00 PM Central — email first and we'll find a time.