We have recently moved to our brand new office in Paddington and with multiple engineering teams being based here we would love to show you the great stuff we develop here. We're working on improving algorithms behind search, building planet-wide telemetry platform processing up to 10mln of messages per second, making mobile typing smarter and more...
We're not going to lie, with so many projects going on we need more talent so this will be a chance for you to learn about whether we could be an inetersting company for you.
Oh, and there will be refreshments… don’t worry, we know how to treat our guests :)
TOPIC: "How to build a web-scale query auto-completion system."
In this talk, I will give an overview of Bing’s Query Auto Complete system focusing on the engineering and the machine learning challenges to build such a web scale low-latency system.
Abhishek is a Principal Scientist at Microsoft and technical lead of Bing’s query auto complete feature. He has been working at Microsoft since 2010 and his previous projects have included Bing News, Bing Speller and Bing Related Searches. Abhishek is passionate about applying statistical modelling to tackle complex search problems. He holds a PhD from Edinburgh University’s Informatics school where his dissertation on Statistical Machine Translation was supervised by Phillip Koehn.
TOPIC: "How to theme a keyboard."
SwiftKey enables multilingual typing on Android phones around the world. With support for over 150 languages, we have thousands of different keyboard layout configurations that our theming system has to handle. In this talk I’ll explain the lessons we’ve learnt whilst building a theming system capable of adapting and growing to new text input methods whilst still being highly performant.
Olli is a software engineer on the Android IME team in SwiftKey. His projects include building emoji prediction models, analysing performance bottlenecks, and reducing some of them too.
TOPIC “Aria: A platform for Real-time analytics at scale”
ARIA is a multi-tenant real-time data platform within Microsoft's AI and Research division, collecting data from products such as Office, Bing, Skype, Cortana, MSN and more. It processes 10 million events/sec, from more than 800 million devices and services, summing up to around 50 petabytes per month. The platform provides both real-time visualizations, fast adhoc queries, alerting, funnels, experimentation and more all in one place. Our aim is to democratise ingestion, real time aggregation at scale in a fully self-serve manner enabling our customer to easily get insights from their app/services. Building and maintaining a system of this scale requires both an architecture tailored for high performance and a relentless focus on reliability. This talk will cover both the evolution of the platform to the scale it has today and the architecture and strategies that enable its performance.
Luca is a software engineer at Microsoft. He holds a Master degree in Telecommunications Engineering, for the past two years he has been working in ARIA, focusing in real time aggregation and query performance.
Jon worked previously as a lead graphics programmer at games studios such as Atari, EA, Lionhead and Codemaster. This has gave him a lot of experience at designing and implementing a number of rendering engines and tool chains, and dealing with a lot of the difficulties that come from creating 3D content and rendering along the way. He has also worked extensively on procedural model creation.
In addition to the games industry he has also worked in for The Foundry, the VFX software developer, and at design art studios where he designed and built electronics as well as developing embedded systems and high-end rendering techniques for realistic lighting and shading.
At Lift London he has been responsible for researching and developing the 3D texturing and modelling engine for Paint.