Nov 2018 — Aug 2023 · 4 yrs 10 mos

Fossil Group

Nearly five years at Fossil Group, growing from Data Engineer to leading the data team. The Data Platform started on a Node.js data-collection platform on AWS — Kinesis into S3, Lambda transforms, Athena and Redshift — and we spent the next few years making it faster, cheaper, and more reliable: Kafka and Spark on Kubernetes, a full rewrite in Rust, and a move to ClickHouse. Costs came down by more than half along the way. A good place, a good team, and the chapter where I learned to scale data — and to lead — for real.

Senior Data Engineer

Jan 2020 — Aug 2023

Data leader for a team of eight — five engineers and three analysts. Owned the platform end to end, drove the re-platforming for scale and cost, and cut Data Platform expenses by 55%. Automated deployments using Kubernetes (k8s) and Helm charts. Maintained a suite of data tools including Apache Airflow, Datahub, Spark UI, and built Fossil Insights for internal metrics tracking.

Data Engineer

Nov 2018 — Jan 2020

Built and ran the core data platform on AWS, and proposed the data warehouse that became the foundation for product insights. Developed and maintained data visualization dashboards, transitioning from Redash and Google Data Studio to Tableau.

How the platform evolved
  1. Collection on AWS

    Node.js collectors streaming through Kinesis into S3, Lambda transforming data between zones, and serving queries from Athena and Redshift.

  2. Data visualization evolution

    Set up the initial reporting layer using Redash and Google Data Studio, later migrating to Tableau for unified enterprise analytics.

  3. Streaming & storage upgrades

    Moved ingestion from Kinesis Firehose to Kafka, and retuned data types for far better compression — real storage and cost savings.

  4. Compute on Kubernetes (k8s)

    Migrated batch workers onto EKS, deploying and managing applications with Helm charts. Orchestrated workflows using Apache Airflow and cataloged data assets with Datahub.

  5. Platform visibility & Spark

    Processed large-scale datasets with Spark, monitoring job performance with Spark UI, and launched Fossil Insights (our internal monitoring portal).

  6. Rewritten in Rust

    Re-built the platform end to end in Rust for another massive cost cut — faster, leaner, and cheaper to run.

  7. Redshift → ClickHouse

    Swapped the warehouse to ClickHouse on Kubernetes for fast serving, with Next.js dashboards on top.

  8. A second platform on GCP

    Stood up a parallel data platform on GCP to meet specific requirements.

  9. Leading the team

    Grew from engineer into leading the data team — hiring, mentoring, and setting the technical direction.