Project Vigor - built a chat-driven AI coach that transforms personal goals into step-by-step schedules, with a direct iCal integration.
Shipped several LLM-powered tools, including a Slackbot that lets infrastructure and dev teams interact with their AWS environment through conversation.
Designed and built an end-to-end MLOps platform on Kubernetes from scratch: data collection and versioning, GPU training and benchmarking, and model publishing to inference endpoints for vision and natural language workloads.
Built a web-based coloring-book generator and experimented with vector databases and LLM agents.
Mentor engineers on reliability, scalability, and rapid experimentation across multiple teams.
2017 - 2023: Scaling and Leadership
Took a company’s infrastructure from early startup foundations on a default VPC to a SOC 2 Type II certified production platform serving high-throughput OLTP workloads, supporting growth from 20 to 70+ engineers.
Designed self-service platform tooling adopted across 5 product teams shipping multi-tenant SaaS, single-tenant SaaS, and on-prem. A single platform cut deploy times and lifecycle overhead across every delivery model.
Created per-pull-request cloud development environments by extending the team’s existing tooling instead of replacing it. Developer-owned from day one, scaled to 200+ concurrent environments.
Built the Terraform module library, Atlantis workflows, and infrastructure-as-code practices for the broader AWS footprint, including a global networking mesh for disaster recovery and standby failover.
Deployed full observability across Grafana, Prometheus, Loki, and OpenTelemetry, and wrote custom Go services for SRE dashboards, SLO tracking, and incident response workflows.
Led a lean SRE team managing 5 products with 24x7 on-call, maintaining 99.95% uptime and cutting mean time to recovery by 40%.
Built a global email security validation framework spanning 12 AWS regions and 100+ domains, using a neural network to generate content and measure how well email gateways caught it.
Embedded with a Fortune 100 client managing tens of thousands of servers across 100-200 AWS accounts. Designed serverless map-reduce pipelines for IoT telemetry.
Architected global cloud deployments for multiple clients, both active-active and active-passive failover.
Designed and built a network intrusion-detection and hunt tool end to end: sensors in Go and Rust, the backend and analytics layer, and a React component library and full UI from scratch.
Built a Modbus ICS simulation modeling a complete town water-treatment system, used to test and train against industrial-control-system threats.
Founded a data company, architected a platform processing 350M geospatial records per day with a custom S3-based trie index over ~200TB, queryable on multiple dimensions for under $5K/month with self-healing resharding.
Founded a drone company. Designed a modular 3D-printable payload ecosystem for commercial drones with data/power rail spine and clip-in sensor modules. Successful flight tests across multiple airframes.
Built and led a growth-focused sales-engineering practice serving finance and energy sectors, shrinking proof-of-value cycles from weeks to days.
Guided cross-functional teams through rapid prototyping, turning data insights into customer-facing features at startup speed.
2015 - 2017: Product Development and Delivery
Built an OLAP analytics platform (Kafka, Spark, Elasticsearch) serving multiple federal agencies from a single framework with a lean engineering team of under 10.
Had the privilege of working with Dr. Brad Cox, creator of Objective-C, building systems to deliver his Spark-based analytics at scale.
First infrastructure-at-scale role. Managed hybrid infrastructure between physical server racks and AWS, learning horizontal scaling in a big-data context. Foundation for everything that came after.
2011 - 2015: Big-Data Engineering and Analysis
Designed and deployed a visual workflow tool that automated complex multi-source analysis, scaling from zero to 5,000+ users and doubling the platform’s addressable market.
Built a natural language report triage system that tokenized intelligence reports, trained labels with volunteer analysts, and automatically sorted incoming reports by priority. Replaced a manual process that was never complete with one that was 80-90% accurate on its worst day.
Led big-data solutions using Hadoop/Solr stacks to support global military and intelligence missions.
2006 - 2011: Building Engineering Foundations
USAF Cryptologic Linguist stationed at NSA Hawaii. Took raw intelligence feeds across multiple data sources and built conclusions about targets.
Designed a multi-source intelligence correlation process that gave geolocation to data that previously had none. A separate team (3-4 people) was formed around this capability.
Automated a manual cryptographic process by building a distributed compute system across office workstations that pre-decrypted messages overnight.