Kyler Liu

I run Tencent Content Platform, which aims to build encompassing content experience for all the consumer apps e.g video, feeds and news etc. The key pillar comprises content creation, content publish, content process, content intelligence and content incentive. We strive to bring rich, diversified and original content into Tencent PCG (Platform & Content Group) app family which touches hundreds of millions of end users. Our state-of-art content technology offers content creators (PGC, UGC, Business and institutional account, etc.) variety of capabilities engage with consumers, promote brand awareness and realize monetization. We particularly focus on original content creation experience to bolster a healthy, diverse and vibrant eco-system.

Bring equality and improve humanity on medical treatment and healthcare with technology innovation, we use cutting edge technology data mining, computer vision, machine learning, to weave engineering into research (theory) so they work together to build products (in practice) that improve the quality of life. - (Feb 2019): I run Product, Data, Applied Machine Learning, Design at Tencent Healthcare. Our app Tencent Health is running on Wechat wallet as 1st party healthcare offering that touches 1 billion users, we provide online doctor consultations (with ML enabled triage, pre-diagnosis prior to patient& doctor meeting), prescription delivery, vaccine appointment, and medipedia search & discovery etc. We also partner with hospital on auxiliary diagnosis using AI medical image reading to improve the efficacy in a resource deprived setting. - (July 2018): Head of Data, and Applied Machine Learning, Tencent Health

Head of Product, responsible for Uber Risk & Safety Platform, Knowledge Graph, and Device Intelligence - (April 2017): Founding member to lead ML based knowledge graph to solve Safety, Risk and Growth challenge in a real time double-sided marketplace Build unified Device Intelligence product for Uber across all the product lines mobile (iOS, Android), web (m.uber.com, uberEATS.com, etc) and 3rd party (Google maps, Facebook messenger, etc) - (June 2016): Build real-time ML product to tackle Safety (driving safety, NLP based ticket classification, interpersonal safety experience, autonomous) and Risk problem (financial fraud detection and marketplace incentive fraud), we apply multi-faceted machine learning algo on behavioral data, telematics, GPS to create trustworthy consumer mobile experience in real world

Delivered an integrated set of large-scale, high-dimensional Customer Knowledge big data Graph and enabled Machine Learning based predictive analytics to drive Consumer business (engagement and monetization) - Drove cross-functional adtech product (marketing, product, engineering, science, and designer) to set qualitative objectives and KPIs - Integrated Segment Designer for Windows, Office, Xbox and MSN consumer marketing campaigns to generate user segments for Windows 10 in-prod targeting - Led transformation of Customer Knowledge Graph from concept into a scalable platform spanning across Microsoft Consumer and Commercial business, with billions of nodes & hundreds of billions of edges, to model Microsoft customers, devices, products, applications, subscriptions, payment instruments, offers, usage, family service etc. - Conceived, designed and launched Segmentation-as-a-service, which is the V1 version of Windows 10 in-prod customer targeting capability, this enables first party targeting delivering billions of impressions each month across multiple Windows 10 canvases Start menu, http://account.microsoft.com/, Lock screen, Edge etc. - Led the Machine Learning predictive analytics with data science and engineering to drive targeting lift (100~150% CTR and app launch conversion) compared with human defined segments. The model we use are semi-supervised learning and collaborative filtering - Planned and led the development of Machine Learning models to discover insights, Customer engagement score, Customer influence score, Expected Revenue 12month, Churn prediction (Office 365) to drive cross-sell and up-sell scenario

Commerce platform knowledge service team including big data, Payment, Risk and anti-fraud and machine learning solution in support of Microsoft's largest scale, most business critical cloud services. Led Online Commerce Anti-Fraud service modernization effort to optimize the experience for good customer while blocked the fraud attack using multi-tiered Machine Learning based online service, this supports all the Microsoft consumer and commercial multi-billion dollar online businesses including Xbox Marketplace, Windows Store, MicrosoftStore.com, Bing Rewards, Office 365 and Azure - Drove cross-functional team to set qualitative objectives and quantitative goals - Envisioned, simplified and launched modern Anti-Fraud service integrating with multiple Commerce stacks (Ordering, Billing, Payment, Chargeback, Provisioning, etc.) by delivering Service-oriented architecture REST API - Built Machine Learning as opposed to rule-based Risk service aiming to creating a friction free experience for good customers, reducing Risk engine decision false positive and false negative - Designed and integrated Risk Machine Learning random forest model into Online shadow environment for better model iteration, feature development and label collection - Prototyped, analyzed real-time Risk decision telemetry (Approval, Reject, Manual review, Challenge etc.) to discover and iterate next feature requirement

Built v1 Service Intelligence solution for Microsoft Azure service, the platform later grew into one of 6 key service vectors of SQL Server Business Intelligence group

Built web analytic data pipeline and customer intelligence for Microsoft.com (#1 commercial website in UV/PV, top 5 largest web site world wide in any kind)

Ran the enterprise DW and BI solution for General Motor analytic platform

Built/ supported one of largest online marketplace and distributed transaction platform world wide, delivered a variety of state-of-art end-to-end web analytic,Business Intelligence and DW solutions for eBay marketplace