Marketplace Data Observability team. We manage pipelines for processing large volumes of business-critical ads data and powering analytics.

Technical domain includes

* Distributed Systems

* Batch/Stream Processing

* Logging

* Data Schematization, Modeling, Observability, & Warehousing

* Job Orchestration

Business domain includes

* Ad Metrics

* Ad Attribution

* Advertiser Billing

* Ad Delivery Monetization

* Bot Filtering

Mentorship: I've guided many engineers (NA + EU) during their onboarding, projects (design/scoping/prioritization), and oncall shifts.

Incident response: We're one of the most oncall-heavy teams at Yelp, and I've been responsible for accurately assessing the (financial) impact of many incidents, quickly resolving them, writing postmortems, proposing platform/infra improvements, and making our systems more failure-resilient.

Team: Ads Backend

Simplified campaign management for advertisers with multiple budgets (i.e. regular + promotional), allowing bids to be set more efficiently:

* Created a new MySQL table for representing ad campaign budget changes and a Cassandra table for representing an ad campaign's spend within a specific budget period

* Created APIs for querying/updating the above tables (i.e. performing budget edits, getting spend info, etc)

Languages: Python

Team: Ads Backend

Developed a package for logging user interactions with Yelp ads (impressions and clicks) so that new ad products could be monetized quickly and safely:

* Implemented schema validation for heavily nested revenue-critical logs, thereby reducing the risk of malformed data entering Yelp’s ads systems

* Ran performance tests to ensure that the package had a tolerable impact on page load times

* Rolled the package out to 100% of ad traffic and added SignalFx metrics reporting to make it easier to detect logging issues in the future

* Used Spark to compare data produced by the package against data produced by the legacy system

Languages: Python

Team: Game Services

Developed an API for searching usernames and retrieving player info within popular games:

* Implemented a new endpoint that allows you to retrieve information about a player by supplying either an exact or partial name

* Utilized an Elasticsearch-backed service for the partial-matching aspect of this endpoint, and created an SQS workflow for indexing player data in Elasticsearch

* Performed various performance tests using JMeter and optimized reads/writes

Made performance improvements to a graph-based service that managed player relationships:

* Modularized SQS consumers and redeployed them with separate autoscaling

Languages: Go, Java

Team: Video Player & Dashboard

Made several improvements to Vidyard's video player, focusing on user experience and metrics reporting:

* Simplified code for video playback, reducing load times on desktop and mobile

* Identified and fixed issues where video engagement data wasn't being properly reported during video playback

* Refactored code that controlled the player's responsiveness under different screen resolutions, improving the user experience on mobile

* Improved the player's accessibility, localization, and cross-device/browser compatibility

* Redesigned existing player components and expanded customization options

Languages: JavaScript, CSS/HTML, and Ruby on Rails