Can Envarli

Group manager for the Media Infrastructure Engineering team. The group consists of 5 teams and over 50 people across four hubs (San Francisco, Boston, London, Seattle) + various remote locations, and owns the following responsibilities for all media on Twitter and Periscope: * Media ingestion, transcoding, CDN origin services * Client Infrastructure for media consumption, media uploads, and live video broadcasting * Video Data Infrastructure - Full-stack ownership of video player telemetry, data collection, aggregation, and certification * Client Events Infrastructure - Common client infrastructure for emitting telemetry at Twitter mobile and web applications

Manager of the Media Platform team. The team consists of 10 software engineers and is responsible for ingestion, transcoding, and CDN origin for all images, videos on demand, and animated gifs on Twitter. Under my leadership, the team: * Recovered from attrition of two most senior engineers on the team and the engineering manager, with zero attrition to date, and hired 3 engineers to backfill all attrition. * Improved the on-call experience significantly, reducing alert counts by over 60%. * Improved video playback performance significantly for our customers, reducing video file sizes by 10% to 30% without compromising picture quality. * Solidified our custom A/B experimentation framework our methodology, rapidly accelerating our pace of experimentation

* Member of the team that implemented the V1 backend for video ingestion and transcoding for native video on Twitter. * Implemented analytics stack for video publishing metrics after the V1 launch. Media Platform: Shipped V1 of consumer/native video support on Twitter. Worked on fun and challenging problems such as live streaming of NFL games to millions of viewers with targeted ad breaks. * Tech lead and key contributor on scaling up video processing stack for longer videos (30s -> 10m) and migrating/eliminating other video processing stacks. * Designed and implemented a backend API for related video suggestions using collaborative filtering and search results, leveraging existing tech stack. * Tech lead + key contributor for V1 (using http redirects without discontinuity in timestamps) + V2 (custom per-playback HLS manifests with discontinuity in timestamps) implementations for server side targeted ad insertion for live streaming. * Participated in ~40 interview loops. Member of centralized hiring committee. Member of interviewing shepherds group. * Performed duties as acting manager three times under two different managers, one of which was seven weeks long. Selected as team tech lead after the existing tech lead left the team.

Managed a team of 5-8 software development engineers in the Display Ads Marketplace team in the areas of Allocation/Availability/Pricing. During my leadership, the team: * Rewrote the allocation optimization system for scale and optimality (delivery simulation-based, distributed job) while continuing to support existing optimization system. * Reduced optimization cycle to 2 hours from 24 hours, and raised coverage of demand and supply to 100% from 25%. * Rewrote the booking system for scale, speed, and optimality (delivery-simulation based, distributed cache) while continuing to support and make performance improvements to existing booking system. 90th percentile response time reduced to 5s from 120s in the legacy system, and 1s in the new system. Went through multiple combined engineering transitions (Dev/Test, Dev/Test/Ops). Extended the yield management system to accommodate 3rd party exchange supply for guaranteed display ads booking. * Achieved MSPoll WHI scores of 85 (2013) and 88 (2014). Drove recruiting for the team (3 hires) and participated in ~50 interview loops for the team and the larger organization. * Selected to Microsoft High Potential Leadership Program FY 2013 & 2014.

Implemented the first global inventory optimization system for the Microsoft Display Ads Technology stack based on a linear programming solution. The optimization impacts a business that brings in ~$1B in annual revenue. • Collaborated with Microsoft Research on the initial algorithm and proved the concept with a prototype • Independently designed and implemented various dimensionality reduction algorithms to achieve a working solution in production. Further improved the algorithmic solution by using piecewise linear curves. • Filed five patent applications (one granted). • Designed marketplace and allocation algorithms for combining reserved and unreserved orders. • Implemented various GUIs on my own initiative to gain marketplace insights (used in demos) and for debugging and investigating issues. Implemented various algorithms and performance improvements in the booking and pricing system. Investigated and resolved numerous production issues in booking system. • Selected to Microsoft High Potential Leadership Program FY 2011, 2012

Chief Engineer for RoboCup 2007 Atlanta. Maintained event website. Designed venue layout. Wrote RFPs and assisted in seeking contractors. Venue lead throughout setup, run of show, and teardown. Managed contractors and volunteers throughout event.

Worked on prototype projects for the Display Ads Inventory Management engineering team.