This topic describes how we used and promoted Qoder CN to significantly improve coding efficiency and quality, driving the adoption of AI programming within our team.
My name is Di Jianye, and my alias is Xue Yue. I joined Alibaba in February 2014. My career at Alibaba has progressed through several stages. From 2014 to 2017, I worked on research and development for the 1688 service. In 2018, I joined ICBU (Alibaba.com), where I initially worked on ordering and CRM-related tasks. For the past two years, I have been responsible for the overall merchant architecture and exploring AI innovations.
I spend a lot of time coding. A few years ago, I spent only about 30% of my day coding. However, for the past two years, that has increased to about 70% to 80%.
At the end of 2022, when general-purpose Large Language Models (LLMs) became popular, there was widespread amazement and heavy investment in the technology. I also decided to focus on AI around May 2023.
GitHub Copilot was popular at the time, so we started trying it. However, we stopped using it because of security and cost concerns. I then tried all the Copilot-like tools available within the company, but found that they did not meet my needs. In September 2023, a colleague recommended Qoder CN to me. After using it extensively, I found it to be very effective.
When writing code, a key concern is whether a tool can intelligently generate code. For example, with Qoder CN, I can write one line of code and it completes the rest for me. This changes my workflow to focus on reviewing and accepting suggestions. It greatly reduces fatigue and helps avoid repetitive manual tasks.
Second, when coding, it is impossible to know everything. We might be unfamiliar with certain methods or syntax. I need to find answers quickly during this process. Qoder CN helps me find answers fast, so I do not need to switch back and forth between my editor and a browser. This has greatly improved my programming experience.
Qoder CN generates a significant amount of my code. In my experience, Qoder CN accounts for about 30% to 40% of my code.
My favorite feature is its automatic code completion. In the six months since I started using it last September, I have noticed that Qoder CN has become increasingly intelligent. It can now generate code that matches my personal coding style.
Another great feature is unit test generation. The accuracy of the generated unit tests is very high, around 80% to 90%.
The benefits are very clear. The feature set of Qoder CN is becoming more comprehensive. First, it covers almost all the features of other Copilot-like tools on the market. Second, its intelligence is impressive. I frequently use its automatic code generation feature and have found the adoption rate to be very high, around 60% to 70%.
In May 2023, our architecture group was responsible for promoting AI Coding Assistant and introduced AI Coding Assistant to the team. After we promoted AI Coding Assistant, everyone was very impressed. I recommend that you start using it. If you use it for two or three days and are patient with it, you will discover how great AI Coding Assistant is. I also manage an outsourced team and have had them all adopt it. AI programming is undoubtedly the current and future trend.
Customer Cases