Sunil Shah

I manage part of the Systems team, we are the lowest level infrastructure team and look after large scale Kubernetes infrastructure. We are hiring extensively for experience with Kubernetes, accelerators, networking and cloud infrastructure, in the SF Bay Area, Seattle and in London. Please reach out (with a message) if you are interested.

I managed the Compute Lifecycle team, we own systems that automate provisioning of Kubernetes clusters at large scale across all three major cloud providers.

I managed the Compute Infrastructure team at Airbnb, part of the Cloud Infra organisation. We're a team of 16 (~11 SWEs, 4 DevOps engineers, 1 SRE) who maintain the core Kubernetes infrastructure and the service management tooling that sits on top of it called OneTouch. Our clusters and tooling orchestrate just over 50% of Airbnb's compute, including almost all ~thousand production services that comprise the website, as well as a handful of data and ML infra workloads. In 2021, I owned hiring for Cloud Infra, personally interviewing 120 candidates and extending 17 offers. 25% of candidates interviewed were from under-represented backgrounds. I've put 5 engineers on my team up for promotion at levels all the way up to senior staff engineer. Some of the team's achievements so far: + Merged two prior teams allowing us to build a sustainable on-call rotation and reduce technical debt. + Built a culture of regular firedrills to validate on-call readiness. + Conducted multiple pre-mortems to prepare for IPO and scalability events. + Built a performance and efficiency subteam that has contributed $XX millions of AWS cost savings to Airbnb's AWS spend. + Drove adoption of continuous delivery at Airbnb through alerting and monitoring of stale namespaces, thus enabling faster response to security incidents. + Drove Kubernetes to complete adoption for all stateless, non-legacy services. + Built a roadmap, product offering and strategy for stateful services on Kubernetes, onboarding Flink, Presto, and TiDB, as well as taking ownership of existing shadow machine learning infrastructure clusters. + Joined the CNCF End User community and presented at every KubeCon for the last 4 years. Contributed changes to upstream Cluster Autoscaler project. + Drafted a whitepaper and submitted to two academic conferences. + Implemented patch management process and worked closely with InfoSec as part of IPO readiness efforts and beyond.

Engineering Manager in the Data and Production Engineering org at Yelp. I managed the San Francisco side of the 16 person Compute team, which develops and supports the Kubernetes and Mesos based infrastructure used for running batches, integration tests, services and machine learning workloads. Our clusters serve traffic from ~ 70 million MAUs (as of June 2018). Some of the things I'm proudest of from my time at Yelp: + We bootstrapped the Compute team from two other teams, unifying our clustering infrastructure and removing a significant amount of redundant code. + We built a fun, inclusive and collegial team culture. + We built processes to allow the team to run effectively, including overhauling our onboarding documentation, building a quarterly prioritisation process, setting the team up for agile and building a strong and well supported on-call rotation. + We sourced and recruited hundreds of candidates to help grow Yelp's SRE organisation. + We shipped easy to use machine learning infrastructure (Spark on Mesos) that is now managed by a standalone team and used by 20+ teams. + We prototyped integration of Kubernetes and built a two year roadmap for full scale migration from Mesos to it. + We contributed our first patch to the upstream Mesos community. + We overhauled Yelp's critical batch processing system, building support for scheduling containerised batches into it and significantly improved its reliability and robustness. + We gave 16 talks at various conferences including MesosCon (which we keynoted), Velocity and LISA.

Early employee at Mesosphere. I worked on a variety of things since starting! + Worked on the DC/OS SDK. + Managed (engineering and product) a small remote team that built a continuous delivery product (known as Velocity). This involved working with multiple other vendors to tightly integrate their solutions with the Mesosphere DC/OS. We facilitated the Artifactory, GitLab, Nexus, and Spinnaker integrations. + Helped implement a large scale migration to Jenkins on DC/OS (one of the ways in which we dogfood DC/OS amongst the engineering team). + Supported internal developers with test tooling and test infrastructure. + Helped recruit nearly all of Mesosphere's interns and acted as the engineering lead for our internship and university recruiting efforts. + Built out our university recruiting efforts at UC Berkeley. + Built out the backend for our (now deprecated) Mesos cluster provisioning service on Google Compute Engine. + Worked on our Mesos monitoring tool. + Helped build the DC/OS command line interface, the CLI for your datacenter. + Presented on Mesos and Mesosphere's projects 35 times at several meetups and conferences including ApacheCon, LinuxCon and Velocity. + Organised a weekly engineering speaker series. + Miscellaneous contributions to docs and customer support.

Member of the Data Team working on Last.fm's music ingestion system - using Java, Spring, Postgres SQL, MySQL, Hive and Hadoop.

Finished the technology graduate scheme, rotating through two teams. Rotation 1- SOIT Application Management Primary C# front end developer in London developing screens to aid the Securities Operations IT support team. Developed for FISS (Fixed Income Settlements System), a mature settlements system that settles both fixed income and equities trades and used by thousands of users across the bank. Rotation 2- Equity Derivatives - Soft Market Data Development Building out a greenfield market data system in Java that provides a Pricing Service to various internal applications.

Helped draft a proposal sent to the Minister for Human Resources and Development. Set up OLPC India on Salesforce.com, importing thousands of contacts from industry to civil service to government. Demonstrated the laptop to the CSR organisation of KPMG.

Intern in the Arbitrage IT team, evaluated and built a prototype complex event processing / event stream processing tool using the open source NEsper engine. Gained approval for NEsper to be used inside the bank via Technology Council. Approval required evaluation of vendor products, several presentations to the council, setting up a working group with the middleware team and a unanimous 'for' vote by voting members.

Intern in Securitisation IT department, working on the custom application built in-house for the CMBS Origination and Servicing team. Developed an addition to their custom reporting tool to allow easier customisation by users and faster access to commonly used queries without administrator intervention.