Projects

A timeline of key systems, infrastructure builds, and software engineering projects from my career. Click on any project to read a detailed writeup of its design and outcomes.

Personal 2026

chicagobuss.dev Hosting Infrastructure

A lightweight personal hosting stack using a modern Rust-based static site generator, Caddy, and automated cloud deployments.

  • Configured Zola to compile static HTML pages locally in milliseconds.
  • Deployed Caddy on an Oracle Cloud ARM instance with auto Let's Encrypt certificates.
  • Built scripts leveraging Porkbun DNS API and VM iptables for secure traffic management.
Zola Caddy Oracle Cloud ARM Porkbun API bash
Duckbill Mar 2025 - Present

Duckbill Data Pipelines

Overseeing data infrastructure, security, and developer velocity for 0 -> 1 product development.

  • Architected and built a low-overhead data pipeline using S3, Parquet, DuckDB, and Clickhouse.
  • Setup Clickhouse Cloud Private Links and private Tailnets to secure workstations.
  • Guided engineering adoption of LLM-based agentic workflows to speed up launch milestones.
DuckDB Clickhouse S3 Tailscale Python LLM Tooling
TRM Labs Jan 2024 - Feb 2025

TRM Multi-Cloud Orchestration

Led multi-cloud orchestration, FedRAMP buildouts, and infrastructure cost optimization.

  • Designed and deployed a large-scale, non-Kubernetes multi-cloud processing platform.
  • Optimized Airflow performance and reduced data engineering IaC footprint.
  • Served as Kubernetes SME to lead FedRAMP compliance and container networking audit readiness.
Kubernetes Airflow Terraform Multi-Cloud FedRAMP
Pinecone Oct 2022 - May 2023

Pinecone Pulumi IaC & Zero-Trust

Refactored next-gen Pinecone platform IaC, shrinking code footprint 10x and implementing zero-trust access.

  • Shrunk Pulumi infrastructure-as-code footprint by 10x using Python.
  • Removed public IP exposure for CloudSQL databases and setup static egress IPs.
  • Transitioned from long-lived credentials to GCP Workload Identity and Flux CD GitOps.
Pulumi Python Kubernetes Flux GCP Workload Identity
Confluent 2021 - 2022

Confluent Vault Security Hardening

Pioneered Vault dynamic secrets integration for Go and Python apps to eliminate static cloud credentials.

  • Integrated HashiCorp Vault dynamic secrets for ephemeral cross-cloud authentication.
  • Eliminated static long-lived credentials across Go and Python cloud microservices.
  • Led engineering compliance efforts to secure SOC2, PCI, and ISO27001 certifications.
HashiCorp Vault Dynamic Secrets Go Python Security Compliance
Confluent 2019 - 2022

Confluent Cloud Margin Optimization

Founded the cloud cost engineering function, managing hundreds of millions in commitments and scaling cloud gross margins from negative to 70%.

  • Scaled Confluent Cloud gross margins from -50% to +70% via structural optimization.
  • Managed over $100M+ in multi-cloud commitments including AWS Savings Plans and GCP CUDs.
  • Drove Kubernetes container bin-packing initiatives, saving $600k+ in monthly recurring spend.
FinOps AWS GCP Kubernetes RI Management
Confluent Summer 2018

Confluent Customer Metrics Pipeline

Led architectural design and rollout for a cloud-agnostic metrics API, delivering cluster statistics to customers.

  • Co-designed a cloud-agnostic, highly-available telemetry ingestion pipeline for user clusters.
  • Constructed Confluent's first internal billing and cost data warehouse on BigQuery.
  • Provided real-time metrics back to product, UI, and external developer dashboards.
Kafka BigQuery Data Pipelines Go Python
Citadel Spring 2017

Citadel DataKeep Secure Access

Built DataKeep, a secure S3 proxying gateway mapping Active Directory identities to temporary IAM credentials.

  • Designed DataKeep to proxy and audit S3 data access under active directory authentication.
  • Generated dynamic temporary IAM credentials to eliminate static developer access keys.
  • Structured private VPC layouts and secure gateway paths for hybrid cloud environments.
AWS IAM Active Directory S3 Proxies Go
Citadel Summer - Winter 2016

Citadel BSync Replication

Built a fault-tolerant, distributed data synchronization application managing on-demand transfers to the cloud.

  • Built BSync, a distributed data synchronization tool to copy datasets to and from AWS.
  • Leveraged HashiCorp Consul for distributed coordination and cluster locking.
  • Managed the ingestion and cataloging of dozens of incoming vendor datasets weekly.
Consul Distributed Locks HTTP API Python Go
Signal Fall 2014 - Spring 2015

Signal Stats Pipeline Performance

Built a Python-based pre-aggregator that bypassed Storm bottlenecks, speeding up statistic ingestion 50x.

  • Engineered a Python Kafka pre-aggregator that improved ingestion efficiency by 50x.
  • Designed a new stats stack featuring Netflix Suro, Apache Kafka, Apache Storm, and KairosDB.
  • Optimized core configuration database latencies, cutting page load times by 20x.
Kafka Storm MongoDB KairosDB Python
Signal 2011 - 2013

Signal Zerg Cloud Abstraction

Built Zerg, a custom multi-cloud abstraction layer automating VM provisioning and DNS management across AWS and OpenStack.

  • Authored Zerg, a multi-cloud Flask UI/API using Libcloud to manage server instances.
  • Configured Git-controlled Puppet repositories to unify configuration across AWS, OpenStack, and local VMs.
  • Implemented local development sandboxing on VirtualBox using mirrored production manifests.
Python Flask Libcloud Puppet VirtualBox
IMC 2010

IMC Overclocked Server R&D

Researched and built liquid-cooled, overclocked rackmount systems to maximize trade execution speeds.

  • Designed custom overclocked server architectures, achieving a 30% performance boost.
  • Reduced datacenter colocation space footprint by 50% through high-density chassis layouts.
  • Tuned Linux kernels, thread pin affinities, and JVM options to isolate trading execution loops.
Low-Latency Tuning Hardware R&D Overclocking OS Thread-Pinning
Orbitz 2008 - 2009

Orbitz Hourly NOC Checklist Tool

Built a custom PHP/MySQL web application replacing manual NOC spreadsheets, remaining in use for over a decade.

  • Designed a PHP/MySQL NOC checklist application to replace fragile spreadsheet logs.
  • Configured checklists using structured XML schema files for flexible updates.
  • Integrated the Graphite API to display dozens of status charts inline on a single page.
PHP MySQL XML Graphite API Web Development