The Microsoft Turing team is an innovative research and engineering team working on state-of-the-art deep learning models to power end user experiences across the company. Applied scientists and software engineers in the Turing team work together to push the boundaries of large-scale deep learning models on Turing’s world class distributed training infrastructure.
The team works on projects across the entire deep learning life-cycle. This includes deep learning models for language representation, language generation, image representations, and reinforcement learning which are then leveraged for down stream tasks for Bing, Office, Xbox, and more. Our Turing NLG (natural language generation) model is 17 billion parameters and powers many generative models within Microsoft. The Turing ULR (universal language representation) model is currently the top model on the XTREME leaderboard. Additionally, we work on large-scale distributed training, improving training performance, and run-time optimized inference allowing us to utilize the power of large models in production scenarios with tight latency constraints. The scenarios our models power reach hundreds of millions of users every month and span across natural language processing, computer vision, video processing, reinforcement learning, and recommendation systems.
In addition to ensuring that Microsoft is pushing the state-of-the-art large-scale models, the Turing team also works on improving experiences across Microsoft’s products and services by working on high value scenarios. The Turing team has shipped capabilities across all of many Microsoft scenarios including AutoSuggest, Microsoft Word’s new find feature, quick reply on Outlook, Smart Compose features in Word, image enhancement models, recommendations, as well as many other features across Office, Bing, and Xbox. Additionally, we are working on a few entirely new experiences such as semantic search for enterprise and knowledge management (aka Cortex) for our SharePoint customers.
Are you looking for opportunities to work on state-of-the-art deep learning models and work on impactful features that will be used by 100s of millions of people? Do you want to work with world class GPU training infrastructure to solve real world problems? Would you like to be a part of a diverse, passionate, open-cultured, high visibility team with ambitious goals and a track record for delivering high value impact across the company?
We are hiring talented, highly motivated applied scientists with expertise in deep learning, machine learning, natural language processing, information retrieval, computer vision, video processing, reinforcement learning, data mining, and large-scale distributed computation. We are hiring at all levels from entry level to principal. Our goal is to build large, general models that can work across many downstream tasks delighting millions of users with much improved experiences across all Microsoft products.Responsibilities
As part of the team, you would be driving projects through their entire life-cycle from idea creation through implementation, experimentation and finally shipping to our users. We are looking for strong, motivated, results-oriented applied scientists and software engineers to help drive the design and implementation of the next wave of improvements of products for Bing and Microsoft. You would be working with truly big data by training and extending the capability of deep learning models to provide compelling experiences across our products. Projects can span across modeling, data engineering, model optimization, large scale experimentation, and shipment of cutting edge technology for 100s of millions of users. Due to the wide range of project opportunities, projects are assigned based on candidate's interest and background. Also, the project assignments are fairly permeable. So, one can move around and play different roles depending on their interest.Qualifications
4+ years of experience as an applied science engineer/leader designing, implementing and shipping and models and experiences to customers.
2+ years of experience driving technical and architectural dependencies in the field of machine learning/deep learning.
2+ years of experience with distributed systems delivering large-scale systems into production.
8+ years of programing experience in C++, Python or other similar languages.
Exceptional software development design, debugging and troubleshooting and problem solving skills.
Ability to generate clarity in ambiguous situations.
Proven track record of working across teams to deliver shared solutions to complex problems.
Passion for and ability to quickly understand emerging technologies.
Strong communication and collaboration skills.
BS or MS degree in Computer Science, Mathematics or related fields.
Background in machine learning/deep learning (strongly preferred).
Ability in self-learning, entering new domain, managing through uncertainty in an innovative team environment
Passionate and self-motivated.
Ability and effectiveness working in a significant technical problem domain, in the term of plan, design, execution, continuous release and service operation.
Microsoft is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to age, ancestry, color, family or medical care leave, gender identity or expression, genetic information, marital status, medical condition, national origin, physical or mental disability, political affiliation, protected veteran status, race, religion, sex (including pregnancy), sexual orientation, or any other characteristic protected by applicable laws, regulations and ordinances. We also consider qualified applicants regardless of criminal histories, consistent with legal requirements. If you need assistance and/or a reasonable accommodation due to a disability during the application or the recruiting process, please send a request via the Accommodation request form.
Benefits/perks listed below may vary depending on the nature of your employment with Microsoft and the country where you work.
#search# #webxt# #msturingjobs# #machinelearning#