Jon Eisen

Staff Software Engineer and Technical Lead

I love to build useful software. I’m an experienced polyglot generalist programmer who has built everything from high performance embedded systems to high scalability distributed systems, working solo and leading cross-functional teams across many industries.

For the past decade, I’ve focused on data engineering. My guiding principle is that we must understand what the data represents at a low level to understand what it’s implying at a high level.

Functional programming is my preferred style. Defensive design, exhaustive testing, extensive automation, and self-documenting APIs are all habits of mine. I think a programmer’s greatest skill is writing readable code.

Work History

Data Engineering

2016 - Present

Apple | Speech Recognition

Oct 2023 to Present
Lex - Your Speech Data AI Analyst

Built an AI chatbot with tool access to Speech Catalog and Trino tables, with built-in knowledge harvested from documentation. Enables users to materialize datasets on-demand using SQL transpiling and Spark. Revolutionized accessibility to data across the organization.

Python Interlinked Conductor Trino Spark Kubernetes
Horizon Data Tables

Designed schema and built consolidation pipeline to unify all Speech datasets into shared tables. Developed vendor data onboarding process and machine annotation pipeline to standardize disparate dataset schemas.

Python Trino Bolt Spark
Iceberg Migration

Led an organization-wide effort to migrate 20,000+ datasets from parquet to Iceberg. Built tools, documented process, educated peers, and of course, migrated many datasets. Helped set requirements for tool owners to adopt Iceberg.

Iceberg Parquet Python
Speech Catalog

Designed, built, and maintain a custom Dataset Catalog to manage, discover, and inspect Speech Recognition datasets. Speech Recognition has tens of thousands of datasets for training and evaluation of models. Speech Catalog is enhanced with domain knowledge to help solve challenges managers, engineers, and modelers face on a day-to-day basis.

Python Javascript Trino Iceberg Spark Kubernetes PostgreSQL AWS RDS

Apple | AIML Instrumentation Platform

Dec 2019 to Oct 2023
AIMLInstrumentationStreams

Designed and built an on-device library for inspecting the data that was originally designed to be shipped off-device for later analysis. The library joins together the 1000+ small messages that make up a single Siri request using graph traversal of interconnected components, allowing the caller to easily retrieve information that original came from disparate systems. Now used to enable differential privacy of users by running analyses on-device.

Swift Objective-C
Antero

Designed, built, and maintained a data-in-development inspection and verification tool. Originally built to help understand how the data described a Siri user interaction, it became a vital part of shipping Siri On-Device, accelerating the work of making Siri more efficient and finding issues with data joinability and correctness.

Clojure Javascript Flink Kubernetes ElasticSearch

Activision | Central Tech

Mar 2016 to Dec 2019

Developed and evangelized eSports data ingestion and aggregation system for Call of Duty World League broadcast graphics and analysis. Performed advanced sabremetrics research including win probability modeling. Built Play of the Game simulation matching game code to test weighting changes on real data before production release. Led team of 3 engineers providing operations for databases, CI tooling, deployment platforms, ElasticSearch, and Kafka.

Node.js React Redux Postgres Spark Scala Kafka Kubernetes Helm

Web and Distributed Systems

2012 - 2015

Rafflecopter

Sep 2013 to Oct 2015

Architected and developed core entry processing system with composable middleware, asynchronous queues, and microservices. Redesigned the Rafflecopter Widget, a Javascript embeddable entry form used by thousands of sites for giveaways. Led DevOps infrastructure evolution, migrating from SaltStack to early Kubernetes deployment platform.

Node.js Clojure MongoDB Javascript Clojurescript Reagent SaltStack Kubernetes

Under Armour | eCommerce

July 2012 to Sep 2013

Developed batch-based recommendations engine to replace Omniture on ua.com. Significantly improved page load performance by optimizing recursive JSON rendering system and custom ColdFusion renderer. Built CI/CD deployment system with IRC Chat-Ops for full Continuous Delivery.

Python Disco ColdFusion JSON Javascript Ruby Capistrano

Algorithms and High Performance Computing

2009 - 2012

Northrop Grumman | Electronic Systems

May 2009 to July 2012

Key developer and debugger on Jupiter RADAR system, a modular RADAR running on embedded high performance computing platform. Improved system architecture (UML), intercomponent communication, and CI tooling. Brought SAR (Synthetic Aperture RADAR) algorithm to program completion. Researched High Speed SAS (Synthetic Aperture SONAR) and developed mitigation strategies to achieve consistent resolution, including theory, models, simulation, and high performance implementation.

C++ UML ClearCase MATLAB

Education

Johns Hopkins University Dec 2013
M.S. Applied and Computational Mathematics
Georgia Institute of Technology May 2009
B.S. Applied Mathematics, Highest Honor