The Wikimedia Basis is in search of a Workers Web site Reliability Engineer (SRE) targeted on Machine Studying Infrastructure. You’ll be a part of a distributed group working throughout UTC -5 to UTC +3 (Jap Americas, Europe, and Africa) and report on to the Director of Machine Studying, Chris Albon.
As a Workers SRE specializing in ML infrastructure, your major duty is designing, growing, sustaining, and scaling the foundational infrastructure that allows Wikimedia’s Machine Studying Engineers and Researchers to effectively practice, deploy, and monitor machine studying fashions in manufacturing.
You can be accountable for:
- Designing and implementing sturdy ML infrastructure used for coaching, deployment, monitoring, and scaling of machine studying fashions.
- Bettering reliability, availability, and scalability of ML infrastructure, guaranteeing clean and environment friendly workflows for inside ML engineers and researchers.
- Collaborating carefully with ML engineers, product groups, researchers, SREs, and the Wikimedia volunteer neighborhood to determine infrastructure necessities, resolve operational points, and streamline the ML lifecycle.
- Proactively monitoring and optimizing system efficiency, capability, and safety to take care of excessive service high quality.
- Offering professional steerage and documentation to groups throughout Wikimedia to successfully make the most of the ML infrastructure and greatest practices.
- Mentoring group members and sharing data on infrastructure administration, operational excellence, and reliability engineering.
Abilities and Expertise:
- Candidates ought to be based mostly inside UTC -5 to UTC +3 time zones to make sure good collaboration overlap with the group.
- 7+ years of expertise in Web site Reliability Engineering (SRE), DevOps, or infrastructure engineering roles, with substantial publicity to production-grade machine studying programs.
- Confirmed experience with on-premises infrastructure for machine studying workloads (e.g., Kubernetes, Docker, GPU acceleration, distributed coaching programs).
- Sturdy proficiency with infrastructure automation and configuration administration instruments (e.g., Terraform, Ansible, Helm, Argo CD).
- Expertise implementing observability, monitoring, and logging for ML programs (e.g., Prometheus, Grafana, ELK stack).
- Familiarity with fashionable Python-based ML frameworks (e.g., PyTorch, TensorFlow, scikit-learn).
- Sturdy English communication expertise and luxury working asynchronously throughout international groups.
Qualities which might be vital to us:
- Collaborative, proactive, and independently motivated.
- Skilled working with various, distant groups.
- Dedicated to open-source software program and volunteer communities.
- Systematic thinker targeted on operational excellence and reliability.
Moreover, splendid candidates will excel in no less than considered one of these areas:
- Scalable ML Infrastructure: Deep understanding of scalable infrastructure design for high-performance machine studying coaching and inference workloads.
- Reliability and Operations: Confirmed monitor file guaranteeing excessive reliability and sturdy operations of complicated, distributed ML programs at scale.
- Tooling and Automation: Demonstrated experience creating sturdy tooling and automation options that simplify the deployment, administration, and monitoring of ML infrastructure.
In regards to the Wikimedia Basis
The Wikimedia Basis is the nonprofit group that operates Wikipedia and the opposite Wikimedia free data initiatives. Our imaginative and prescient is a world wherein each single human can freely share within the sum of all data. We consider that everybody has the potential to contribute one thing to our shared data, and that everybody ought to be capable to entry that data freely. We host Wikipedia and the Wikimedia initiatives, construct software program experiences for studying, contributing, and sharing Wikimedia content material, assist the volunteer communities and companions who make Wikimedia attainable, and advocate for insurance policies that allow Wikimedia and free data to thrive.
The Wikimedia Basis is a charitable, not-for-profit group that depends on donations. We obtain donations from hundreds of thousands of people world wide, with a median donation of about $15. We additionally obtain donations by means of institutional grants and items. The Wikimedia Basis is a United States 501(c)(3) tax-exempt group with places of work in San Francisco, California, USA.
As an equal alternative employer, the Wikimedia Basis values having a various workforce and repeatedly strives to take care of an inclusive and equitable office. We encourage folks with a various vary of backgrounds to use. We don’t discriminate towards any individual based mostly upon their race, traits traditionally related to race, faith, coloration, nationwide origin, intercourse, being pregnant or associated medical circumstances, parental standing, sexual orientation, gender identification, gender expression, age, standing as a protected veteran, standing as a person with a incapacity, genetic info, or some other legally protected traits.
The Wikimedia Basis is a remote-first group with workers members together with contractors based mostly 40+ nations*. Salaries on the Wikimedia Basis are set in a method that’s aggressive, equitable, and in line with our values and tradition. The anticipated annual pay vary of this place for candidates based mostly inside the USA is US$129,347 to US$200,824 with a number of individualized components, together with value of residing within the location, being the determinants of the provided pay. For candidates positioned outdoors of the US, the pay vary will likely be adjusted to the nation of rent. We neither ask for nor take into accounts the wage historical past of candidates. The compensation for a profitable applicant will likely be based mostly on their expertise, expertise and placement.
*Please observe that we’re presently in a position to rent within the following nations: Australia, Austria, Bangladesh, Belgium, Brazil, Canada, Colombia, Costa Rica, Croatia, Czech Republic, Denmark, Egypt, Estonia, Finland, France, Germany, Ghana, Greece, India, Indonesia, Eire, Israel, Italy, Kenya, Mexico, Netherlands, Nigeria, Peru, Poland, Singapore, South Africa, Spain, Sweden, Switzerland, Uganda, United Kingdom, United States of America and Uruguay. Our non-US workers are employed by means of an area third occasion Employer of File (EOR).
We periodically evaluate this record to streamline to make sure alignment with our hiring necessities.
All candidates can attain out to their recruiter to know extra concerning the particular pay vary for his or her location in the course of the interview course of.
In case you are a certified applicant requiring help or an lodging to finish any step of the applying course of attributable to a incapacity, chances are you’ll contact us at [email protected] or +1 (415) 839-6885.
Extra info
What does the Wikimedia Foundation do?
What makes Wikipedia different from social media platforms?
News from across the Wikimedia movement