About the job
At Microsoft, The Azure Singularity team is looking for passionate engineers to build the largest deep learning infrastructure service at Microsoft. In this role, you will be tasked with developing new components to bring the latest AI infrastructure innovations to the Singularity platform.
You will collaborate with top engineers within Singularity and on Azure to work on cluster orchestration, job scheduling, containerization, and OS integration. Your work will enable various AI languages and runtimes on Singularity to bring distributed Deep Learning training and inferencing to life.
You will also build infrastructure components required to create, deploy, monitor, and maintain highly available and scalable Microsoft Service Fabric and Kubernetes clusters under your responsibility. You will lead frontline development and customer support, establishing architecture, service excellence policies, and a high quality bar.
Applicants must have a proven record of excellence in engineering and service in a medium to large service organization.
Who we are
We are the engineers at Singularity. We believe that building a world-scale AI supercomputer that solves the fundamental problems of data scientists and AI professionals and takes AI to unprecedented scale is a once-in-a-lifetime opportunity. If you have the same dream as us, join us!
What is Singularity?
AI workloads at scale are always testing the limits of the infrastructure stack. Large scale model training and inferencing with massive amounts of training data on hundredsof thousands of GPUs is a real technical challenge. Singularity is a globally distributed, multi-tenant service that provides a robust, cost-effective, and competitive AI infrastructure (compute, networking, and storage) for AI training and inferencing.
By abstracting workloads from the underlying infrastructure, Singularity creates a common pool of resources that can be dynamically provisioned to take full advantage of expensive GPU computing power, enabling data scientists to productively build, scale, experiment, and iterate on their models based on a robust, high-performance, scalable, and cost-effective distributed infrastructure designed for AI.
At Singularity, we are constantly striving to apply the best ideas from AI, ML, distributed systems, distributed databases, machine learning, information retrieval, networks, and security.
Responsibilities
- Provide a robust container orchestration platform for Singularity
- Design and build the scheduling subsystem responsible for meeting SLAs for AI training and inferencing workloads
- Design and develop control plane APIs for creating and managing client, job, and model metadata
- Provide node management, fault detection, and node repair as a service to improve job/model reliability
- Provide world-class monitoring systems and telemetry pipelines to improve service and job observability for both end-users and operators.
- Codify security and compliance requirements by building and strengthening system defenses against malicious attacks and exploits
- Leverage performance and profiling tools to identify hotspots and bottlenecks across hardware and software boundaries: from CPU, GPU, microcode, OS, networks to product code, and increase end-to-end job performance.
Required Qualifications:
- 1-2 years of programming experience in one of the following languages: Python, C#, Java, C or C++
- Experience working with the Linux operating system and Kubernetes cluster orchestration
- Experience with improving service operations or technical fundamentals
- Excellent collaboration skills
- A master’s degree (or bachelor’s degree with at least 4 years of experience) in computer science or a related field
- At least 5 years of experience developing and deploying production software or services
- Preferred qualifications
- Experience developing in the Kubernetes ecosystem
- Experience using / extending PyTorch / TensorFlow
- Experience developing cloud services, distributed systems, or operating systems at scale
- Experience in GPU (Graphics Processing Units) programming, CUDA /cuDNN/NCCL
AIPLATFORM
This role requires you to meet Microsoft, customer and/or government security requirements. These requirements include, but are not limited to, the following specific security checks: Microsoft Cloud Background Check. This position must pass the Microsoft Cloud Background Check upon hire/transfer and every two years thereafter.
Microsoft is an equal opportunity employer.
In accordance with applicable law, all qualified applicants will be considered for employment without regard to age, ancestry, citizenship, color, family or medical leave, gender identity or expression, genetic information, immigration status, marital status, medical condition, national origin, physical or mental disability, political affiliation, protected veteran or military status, or race, ethnicity, religion, gender (including pregnancy), sexual orientation, or any other characteristic protected by applicable local laws, rules, and regulations.
If you need assistance and/or a reasonable accommodation during the application process because of a disability, read more about requesting accommodations.
About the Company
Every company has a mission. What’s ours We want to empower every person and every company to achieve more. We believe that technology can and should be a force for good, and that meaningful innovation contributes to a better world in the future and today. Our culture not only encourages curiosity, but embraces it.
Every day, we make progress together by showing up as our authentic selves. We show up with a learn-it-all mentality. We cheer others on, knowing that their success doesn’t diminish our own. Every day, we’re willing to recognize our own biases, change our behavior, and embrace differences. When we show up, we achieve more together.
