Areas of Expertise
Designing and evaluating scalable, fault-tolerant backend systems
Batch and streaming data pipelines, data catalogs, data governance
Deep experience with AWS services and large-scale cloud-native systems
Anomaly detection, machine learning-driven monitoring at scale
Driving alignment across engineering, privacy, and product teams
Experience
Netflix Data Platform
May 2022 – Present- Leading the design and development of a unified software and data catalog spanning all technical assets at Netflix, requiring evaluation of competing approaches across metadata management, search, and discovery.
- Architected a data governance platform in partnership with Netflix Privacy team, defining policies that apply across the full breadth of Netflix's technical ecosystem.
Amazon AWS IoT
Dec 2017 – May 2022- Designed and built AWS IoT Device Defender, a publicly launched AWS product announced at AWS re:Invent 2017. The service allows customers to monitor device behaviors and detect anomalous activity using customer-defined rules and machine-learning-based forecasting.
- Evaluated and selected forecasting algorithms to establish behavioral baselines across diverse IoT device populations.
Amazon Amazon Prime
Dec 2011 – Dec 2017- Designed and led the migration of Prime's data platform from batch processing to near real-time aggregation, reducing data freshness from 36 hours to 1 minute. Evaluated architectural trade-offs across Hadoop, Spark, EMR, DynamoDB, S3, and Kinesis.
- Built a generic URL-routing and rule-engine framework for Prime promotional signups, reducing promotion launch time from months to days.
Skills
Open Source & Community
netflix/Metacat – Contributor to Netflix’s federated metadata service for data discovery and management.
dloom – Author and maintainer of an open-source home directory manager, an alternative approach to GNU Stow for managing dotfiles and configurations.
Judging Philosophy
I evaluate projects for practical scalability and value the “why” behind architectural trade-offs – specifically how data integrity and system failures are handled. While creativity is the spark, I look for a foundation of clean architecture and a realistic path toward real-world implementation.