• Designed and oversaw real-time, bandwidth-constrained, tag-based distribution of battlespace objects.

• Wrote an auto-adjusting publish scheduler for the above with a configurable bandwidth usage target.

• Oversaw implementation of my design for a bandwidth-optimized V3 core data replication layer, for rollout to 70+ clients in 4 languages across 12 orgs.

• Shepherded adoption of a V2 core data replication layer to 35+ projects across 7 orgs.

• Benchmark tooling developed for the above later used to track down a year-old hardware regression.

• Forward deployed to the Valiant Shield 2024 wargames, recommending process improvements after.

• Synthesized research papers and industry standards to design a bandwidth-optimized V3 replication layer.

• Designed and implemented the V1 of what would become Anduril's core data replication layer.

• Following initial support from my team, assumed sole ownership of our Data Lake suite of services.

• Created a containerized equivalent to Amazon Athena for SQL queries against offline environments.

• Spearheaded efforts to remove legacy uses of Apache Spark, standardizing SQL under one dialect.

• Redesigned track heatmaps to address RAM and CPU issues. Led to 10x faster large query results.

• Vastly decreased support requests for our time-series database by creating an ergonomic query CLI.

• Wrote a library on top of BadgerDB to solve a critical production issue for a pivotal contract.

• Achieved 10x faster cloud storage deletions, saving the company from costly AWS cluster growth.

• Eliminated wasted traffic on our distributed mesh network via message tag filtering.

• Established comms paths and integrated UCI messaging for an early, major autonomous flight demo.

• Identified months-old latent data corruption in our largest customer account and led repair efforts.

Working as part of the Trading Engineering team, I wrote a novel API for low-latency computing.

This API managed thread pinning and isolation of CPU cores, ultimately enabling faster time-to-trade.