Since the beginning of this year, several major international developer conferences, whether Microsoft Build, Facebook F8 or later Google I/O, have put the banner of “AI first” on the top of the sky.

If this wave of AI is just a few slogans, a few strategies, and a few hot startups, it’s not going to be a big deal. However, what we see is that Google’s major businesses are switching to deep learning, and Microsoft, which is behind the mobile era, has built up an AI team of nearly 10,000. And the situation of domestic first-line big factories, I am afraid is similar.

In the cover report of this issue, we invited experts from sensetime, DuPont, Sonwise, Ciwei, 58.com, Iin Interactive, SciTech, Lulang Software and other AI technology companies to analyze how qualified engineers for various technical positions in the AI field are trained from a practical perspective.

You will learn how to become a:

  • Machine learning algorithm Engineer (Zhang Xiangyu, head of The Algorithm Department of Zhuan Recommendation)
  • Recommended System Engineer (Chen Kaijiang, CTO)
  • Dialogue Systems Engineer (Wu Jinlong, Partner of IIN Interactive Technology)
  • Data Scientist (Hui Lin, DuPont Commercial Data Scientist)
  • Heterogeneous parallel Computing Engineer (Liu Wenzhi, Head of High-performance Computing Department of Sensetime)
  • Speech Recognition Engineer (Chen Xiaoliang, founder of Swise technology)

We will also learn how to choose between the academic approach and the practical approach:

  • Seeking technological Breakthroughs: The Professional path of Deep Learning (Liu Xin, CEO of SciTec)
  • Actual combat path: Advanced methods of machine learning for programmers (Zhi Liang, co-founder of Lulang Software)

information

  • CSDN news
  • A large-scale study of the programming language and code quality at GitHub
  • Why is the biggest Bitcoin mine in China?

Artificial intelligence (ai)

  • Deep learning on iot devices (Tang Jie, Associate Professor, School of Computer Science and Engineering, South China University of Technology)

    Deep learning enables Internet of Things (IoT) devices to parse unstructured multimedia data and intelligently respond to user and environmental events, but with demanding performance and power consumption requirements. The authors explore two ways to successfully integrate deep learning and low-power iot devices.

  • Application of Machine learning in popular Microblog recommendation System (Hou Leiping, Su Chuanjie, Zhu Honglei)

    In recent years, machine learning has made outstanding achievements in search, advertising, recommendation and other fields, and has become one of the most eye-catching technology hot spots. Weibo has also made extensive exploration in machine learning. In the field of recommendation, it has applied machine learning technology to popular weibo, one of the most important products of Weibo, and achieved remarkable improvement.

mobile

  • Interface based message communication decoupling (Peng Fei, 58.com iOS client Architect)

    Code coupling and decoupling is a perennial topic. Coupling is good or bad, only when coupling limits business extension, code reuse and maintenance in a particular business scenario, it is necessary to take some measures to reduce coupling degree. The object-oriented programming world, especially the Java world, has mature decoupling concepts and frameworks, such as IOC and Spring Service. But in objective-C (OC), there is no authoritative framework or tool to deal with code coupling.

  • Retinex Image Enhancement Algorithm and App end transplantation (Zhou Jingjin, Tencent Tiantian P Map Project team)

    Retinex is an image enhancement algorithm commonly used for scenes such as fog removal and night scene enhancement. This paper introduces its implementation and application.

Big data and cloud computing

  • Unit Application to Kubernetes — Migration solution for Microservices Architecture (Leader US, HPE Senior Software Architect)

    Not long ago, Docker announced at the DockerCon held in Copenhagen, Denmark that Docker would actively embrace its old rival Kubernetes, which shocked the industry and further consolidated Kubernetes’ position in the field of container layout. In this paper, the author will focus on the theme of “single application to micro-service architecture migration”. Through specific examples, the author will introduce the process of migration in detail, and bring readers a first-hand solution of great reference value.

  • Construction of Ele. me Big Data Platform (Bi Hongyu, Director of Ele. me Big Data Platform)

    As the demand side of access is more and more diversified, the demand for data use, data storage and computing of big data is also more and more diversified. At the same time, the business is developing rapidly, and the scale of cluster is also expanding rapidly. How to stably support business development through big data platform in such a scenario is not a small challenge. This article will share some experience on toolchain, technology, selection and architecture design of major platforms.

  • Based on the practice of Docker continuous Delivery Platform construction (Liu Xiaoming, Technical director of Operation and Maintenance of Wuagge Steel E-commerce Platform)

    Wuagge, a professional steel service platform jointly established by China Minmetals and Alibaba, brings one-stop purchasing experience to end users by integrating Alibaba’s technological advantages in big data, e-commerce platform and Internet products. This paper is about the exploration and practice of Docker container technology in the process of continuous delivery by the Operation and maintenance technology team of Wuagai. Currently, the release and deployment authority has been opened to the owner of application development to realize the “one-stop” continuous delivery of 7*24 hours, which improves the delivery capacity of the company’s R&D process as a whole.

technology

  • How to identify your users — User aggregation algorithms from different data sources (Yi Zhao, Southbank Software)

    In order to attract users, we have racked our brains with advertising, keyword matching on search engines, and customer outreach through different social software. As the app grows and the user base grows, the question becomes how to identify our “God.” Identification here refers to how to know whether the user registered through different channels is the same person or multiple different users. This article will share the author’s experience and practice in this respect.

  • Interpretation of Ethereum account data storage (Deng Fuxi, System architect of Matrix Technology (Shenzhen) Co., LTD.)

    From the perspective of storage structure, this paper explains the encoding implementation scheme of ethereum’s smart contract and account data model. Through the interpretation of these implementation schemes, we can have a deeper understanding of ethereum’s data model, storage structure, encoding method, consensus principle and other knowledge, which can be used for reference for ethereum-based development and implementation and problem positioning.

  • The difficulty of asynchronous Node.js programming (Huang Dingheng, head of Front-end Node Team at Ele. me)

    For many people, node.js is difficult to develop server-side applications. According to the author, this is not entirely due to Node.js, but more fundamentally due to the difficulty of server-side program development itself. This paper will comprehensively analyze these difficulties from various aspects and share many application development experiences.

  • Chrome extension development – customize HTTP request response header field (3)

    Google Search cannot be added to the IFrame because x-frame-options is added to the HTML Response Header. In order to get rid of X-frame-options, the author tried various solutions and developed the Chrome extension IHeader, which this article focuses on.

subway

  • Sijo Comics – is that how you code today?

Subscribe to the programmers (including iOS, Android, and print), please visit dingyue.programmer.com.cn

Subscribe to consult:

  • Online consultation (QQ) : 2251809102
  • For more information, call 010-64351436
  • For more information, please visit the Programmer editorial Board.