北京招聘

Job # 284783 1048341
Locations China, Beijing
Job families Development (engineering)
Teams Applications and services

Group: Bing! Multimedia Search

Position: Applied Scientist, Senior Applied Scientist, Principal Applied Scientist

Keywords: Computer Vision, Speech Recognition, Web Scale Signal Processing

Work Location: Beijing, China

Reports to: Kit Thambiratnam (alias: kit)



Job Description



Do you want to work on cutting edge computer vision techniques running at web scale? How about the chance to run a vision experiment on 8B images and 1B videos at the press of the button. Imagine the possibilities for deep understanding when you combine hard code vision techniques with user click logs. The Bing Multimedia Team is looking for smart, creative and passionate engineers to join our Multimedia Video Sciences team. We are looking for experts in either computer vision OR speech recognition. The Bing Multimedia team powers image and video search and browse experiences on bing.com.



As part of the team you will work on hard problems such as video deep learning (LSTM, CDNN), video summarization, representative thumbnail selection for a video, image and video action and object annotation, video categorization, duplicate detection and video quality classification. Your work will be used to directly improve ranking of our video search results, providing rich inline preview capabilities for video search results, optimize clicks on video search results and to generate recommended videos. You will be expected to keep abreast of the latest innovations in computer vision and to exploit them for web scale applications. You will be working on vision problems that very few companies can work on because of the scale web indexing requires and the opportunities of cross leveraging user click data to feed into your vision algorithms.



The required qualifications are:



• 3 years+ of computer vision or speech recognition experience. We are hiring at all levels of experience and title and scope will vary accordingly.

• PhD or strong industrial experience in computer vision or speech recognition.

• Favored significant experience in internet-scale engineering systems, search engines, video processing or image processing.

• Favored experience in deep learning techniques for vision or speech recognition

• Committed passion to deep data analysis.

• Strong coding ability. You will primarily work in C++, C# and SCOPE (Internal distributed processing language). Experience in scripting languages such as python, distributed processing such as Hadoop, or set-based languages such as SQL will be favored.

• Ability to clearly communicate and collaborate with team members.

• High level of personal accountability and “self-starter” attitude in order to make progress in ambiguous situations on schedule.



Our developers work on every aspect of the development process including generating feature proposals, designing them, implementing them, shipping them and analyzing data from live production data to improve and innovate further. You will need to be agile, able to jump into any piece of code, learn the latest and best in-house and open source technologies to deliver innovation quickly and with high quality. We also encourage role rotation which means you will have a chance if you wish to try different parts from core engineering to machine learning to user experience development if you choose.



Interested? Contact us immediately to find out how to join.

Bing Multimedia Experiences Overview



Below is a quick overview of some of the experiences the Bing Multimedia empowers globally.



Image Search Experience

Our image search is second-to-none, and we don’t just give you images. We index collections of images from popular sites like Pinterest. We give you related searches for your queries to help you to explore the world of images just by clicking. At the heart of all of this is our ranking technologies that don’t just try to find semantically relevant content, but also consider beauty, size, aspect ratios and watermarking all to bring you simply the best images for your search query.









Play music videos and watch lyrics directly from search results

Now playing your favorite music videos is trivial on bing.com thanks to end to end investments in our music crawling, indexing and semantic conflation technologies. All presented in an immersive carousel experience that lets you jump through all your favorite songs.









Image and Video Search Results in Bing.com

Every time you do a query on Bing.com you are seeing yet another search result powered by the Multimedia team. Our team delivers the image and video results you see in your bing results, on your phone, tablet and desktop.







Video Landing Experience

We bring you the best trending videos, music, TV and movies on our landing experience every day.









Smart Preview

Powered by our video summarization technology, we provide users the capability to quickly preview about 70% of our videos by simply hovering. This helps you better choose exactly which video you want to watch.





Similar Image Search

Powered by our similar search technology based on deep learning techniques, we are able to get similar images to any image you provide. You can see below that most of the cats below are ‘fashionable’ cats since the original image was a fashionable cat.







Video Overlay Experience

We’re giving users the ability to enjoy videos from the entire web all within the comforts of bing.com. Today about 80% of our search results open videos for watching directly within Bing. That means you don’t need to jump across different video sites around the world. Bing simply has one of the most comprehensive video indexes in the world.





Microsoft is an equal opportunity employer and supports workforce diversity.

GCR:CN:DEV:EN

微信分享