Bo Jiang
Algorithm Expert on the ByteHopping Beanbag MarsCode Team
He graduated from the Institute of Computing Technology, Chinese Academy of Sciences with a master's degree, and his research areas include large-scale distributed training and code domain models. As the person in charge of the algorithms of MarsCode Code Completion, Code Completion Pro, Code Q&A and other AI functions, he is responsible for code domain model tuning, inference acceleration, and the incubation of more AI functions for developers. He has long been committed to the combination and practical implementation of soft engineering tasks and big models, and has rich practical experience in the productisation of AI capabilities such as code completion, code generation and code understanding.
Topic
Beanbag MarsCode and the Big Model of Code: realising the future of intelligent programming
With the rapid development of big model technology, code big model is becoming an important force to promote software development and intelligent programming. Beanbag MarsCode, as a programming assistant and IDE product under Beanbag, is committed to applying this cutting-edge technology to actual development by deeply combining with the code big model. In this talk, we will discuss how Beanbag MarsCode uses the code big model to achieve intelligent completion, intelligent prediction, intelligent Q&A and other capabilities, thus improving development efficiency and code quality. We will share the current development status of Beanbag MarsCode product and some explorations and efforts made on the enhancement of code big model capability, and show the application effect of Beanbag MarsCode in different scenarios through specific cases, aiming to provide inspiration and thoughts for the attendees, and look forward to the bright future of intelligent programming together.
