Inside a Chinese SOE: Pay Cuts, Layoffs, and the Case for AI Full-Stack

Hi everyone, I'm Shuangyue. Author of wangEditor, former senior frontend engineer at Baidu and Didi, elite instructor at Muke.com, PMP, and author of Frontend Interview School.

I'm currently running a 'Programmer 1-on-1 Interview Series' where people chat with me 1-on-1 over WeChat voice or text, share their situation and confusion, and I give advice. Below are the main points from the sixth participant (no personal sensitive information), shared for your reference.

Basic info

• Male
• City: Beijing
• Married: No
• Education: Bachelor's degree
• Years working: 4
• Company size: Medium (outsourcing)
• Currently employed: Yes
• Role: Frontend
• Industry: Financial bonds, state-owned enterprise

State-owned enterprises are cutting pay and laying off too

The company's permanent positions are all leadership now. The vast majority, about 90%, of front-line requirements, development, and testing work is outsourced.

When he joined last year, there were still quite a few people. After AI became widespread, many were laid off, and now the company no longer recruits new people.

This year, overtime pay was canceled, bonuses were canceled, and on top of that there are pay cuts and layoffs. Everyone's motivation is low; when there is work, no one is in a hurry to do it, and they slack off ('touch fish') all day.

The projects are still fairly stable and requirements exist; it's just that nobody is in a hurry anymore.

This is also the benefit of doing outsourcing work at an SOE: projects are relatively stable, working hours and intensity are low. At least you're not exhausted, and you have spare time.

AI coding, but no switch to full-stack yet

They started using AI coding this year. The tool they use now is roo-code, and the large model is China's DeepSeek v4.

But they haven't started promoting full-stack development yet. It is still frontend/backend separation, and he is still doing pure frontend. Maybe he has also noticed the trend of AI full-stack, which is why he's worried. My advice is definitely to switch to full-stack, as explained below.

We all know that large traditional organizations like SOEs and banks have fairly old and stable tech stacks and development processes. About two years ago, when I reviewed resumes of bank programmers, they were still using Vue2.

Now, in 2026, this participant says many frontend projects in the company still use Vue2, which was released ten years ago.

The server side is definitely the usual Java + Spring Boot + MySQL all-in-one, another stable technical solution from ten years ago.

Never mind how efficient it is, never mind whether it's new technology, never mind the years of legacy code. It has one advantage: the camp is made of iron while the soldiers come and go. Any programmer can walk in and start doing it.

Move quickly to build full-stack capability

As mentioned above, he is still pure frontend and worried about the future. Actually, the answer is obvious: switch to AI full-stack.

He says he wants to keep doing pure frontend and is looking at remote work for external companies, asking whether that's feasible. This question also came up with a few earlier participants.

Remote work started around 2022. At the time there were two environmental factors: first, the domestic mask period (COVID-19 restrictions), and second, the blockchain development boom. During that period, quite a few people really did find remote jobs, but soon nobody was talking about it. Especially in the last two years, because foreigners also started laying people off and losing jobs, also because AI coding improved efficiency.

Moreover, foreign companies advocated full-stack development much earlier; pure frontend has been gone for a long time. Combined with AI coding, it is all full-stack development now.

You can tell he is somewhat resistant to learning full-stack and backend, so he hesitates and doesn't dare act, always wanting my confirmation. The resistance comes from being a frontend person who has never touched server-side stuff, afraid he won't learn it or do it well.

This mentality is completely understandable. Continuous learning in the programmer profession really is tiring; everyone feels that.

Actually, he has good advantages for learning full-stack now: 1. The job isn't tiring and leaves spare time. 2. The company has real projects he can use as references.

Being unfamiliar just means spending a bit more time and taking a few more detours than others. Even if it's slow and takes half a year to finish learning, that still means next spring. If he needs to find a new job next spring, wouldn't that come in handy? Full-stack is not something you can cram at the last minute.

After talking with me and getting my analysis and confirmation, he has become firm in this direction and hopes to keep at it.