Ding dong! Check out real-world experience from front-line data scientists and big data engineers | IDP Meetup No.02 Review

On April 16th, we held IDP Meetup No.02, and invited front-line data scientists and big data engineers from leading Internet companies - Zhao Xisheng from Tencent and Li Feng from a leading financial technology company, to share from the perspectives of personalized recommendation scenarios and big data processing respectively. Practice and experience of AI development production platform in its daily work. At the same time, Liu Zhe, co-founder and technical director of Baihai Technology, discussed with everyone the programming assistance function and its implementation in IDP.

Dear friends, please check the wonderful content review! (Easter eggs at the end of the article)

Review 1: Practice of Machine Learning Platforms in Personalized Recommendations

Tencent recommended Zhao Xisheng, an architect from China and Taiwan. Taking personalized recommendation as an example, he shared with us how the AI development and production platform makes the training and business application of complex models more efficient. The main contents include:

• Recommend the organizational logic of the middle office
• Data-recommended feature processing: detailing the feature data flow and feature engine architecture
• Training—CTR model training for high-dimensional systems: including typical CTR networks, distributed Gradient Descent, distributed training principles and technical implementation, etc.
• Inference - low-latency, high-availability batch sorting service: including an introduction to the inference engine and specific methods for model optimization
• Business - scalable recommended operation: including recommended usage scenarios, general recommended implementation methods, and recommended link debug.
• IDP's practical experience sharing: Demonstrate the overall process of recommending model building using IDP with actual cases

Live playback: https://www.bilibili.com/video/BV1dY411A76n/

Review 2: Big Data Processing Technology Evolution and Tool Selection

Li Feng, a development engineer of a leading financial technology company, elaborated on the big data technology system and its evolution from the perspective of the development history of big data. Specific content includes:

• Big data processing system
• Data Processing Architecture
• The specific analysis of each module technology and architecture in the big data processing system, including: data source access, data storage, cluster scheduling, general computing, task scheduling, data governance, data security, etc.
• Introduction to the ease of use of IDP in big data processing
  
  

Live playback: https://www.bilibili.com/video/BV1t3411T79Q/

Review 3: Programming-Assisted Implementation in IDP

Liu Zhe, co-founder and technical director of Baihai Technology, elaborated on the specific technology and implementation of programming auxiliary functions that can greatly improve the efficiency of developers. The main contents include:

• The basic composition and core functions of IDE
• What is static analysis and why it is important

• A detailed description of the most important "code completion" features in programming aids

  • The development history of code completion in old editors such as vim/emacs
  • Implementation of code completion in modern IDEs such as IDP Studio/vscode/intelliJ
  • Comparison of various components of code completion
  • Principles, technical difficulties and latest developments of "intelligent completion" assisted by "artificial intelligence"
• Implementation and experience demonstration of code completion in IDP
  
  

Live playback: https://www.bilibili.com/video/BV1ea411v74g/

【Easter egg link】

How to get this sharing PPT: follow the public account Baihai IDP and reply to IDP02 to get it