Leading tech companies are fiercely ramping up AI-related positions in campus recruitment drives, leaving fresh college graduates completely bewildered.

The 2027 autumn campus recruitment is evolving into an AI proficiency assessment.

In early August, ByteDance, Alibaba, and JD.com launched their campus recruitment programs one after another. At ByteDance, technical and product positions account for over 70% of all openings, with new roles including "AI Full-stack Engineer" and "AI Agent Developer"; AI-related positions make up as high as 80% of roles at Alibaba; JD.com introduced AI interviews for the first time... The data is more intuitive: statistics from Maimai show that between January and May 2026, the number of newly released AI positions for campus recruitment rose 47.3% year-on-year, and the penetration rate of AI positions increased from 26.41% to 37.56%. Top tech companies are sending a clear message to fresh graduates with real investment: the future belongs to AI.

But there is still a threshold between "the future belongs to AI" and "I can land a job at a top tech company".

We talked to five graduates from the classes of 2026 and 2027: some submitted 40 to 50 resumes and only got one interview opportunity at a top tech firm; some had internships at leading big tech companies but were rejected by algorithm positions for lack of targeted experience; some pivoted to the AI application layer early and barely caught up with this wave; others tried applying for AI positions last year and got no responses for all their resumes, and finally secured a full-time offer at a big tech company through a non-AI summer internship...

They have different backgrounds and made different choices, but they face the same reality: no matter how high the proportion of AI positions is, it does not mean more opportunities. The threshold is rising, the standards are changing, and in this competition, internship experience has become the most valuable hard currency.

When "being able to use AI" and "understanding AI" are distinguished in job hunting, when a targeted internship experience is more valuable than a prestigious university background, when interviewers keep asking about "workflows", "implementation cases" and "efficiency improvement results", this group of fresh graduates finds that AI is not a shortcut, but another race that requires you to get a head start.

01. Interviewers ask extremely detailed questions, and you will expose your ignorance if you pretend to know what you don't

Frida ｜ Class of 2027 undergraduate, majoring in Chinese Language and Literature at Henan Normal University

Before my senior year even started, I had already submitted 40 to 50 resumes.

I made a spreadsheet to track positions at key companies and my application progress — 15 out of 25 positions I applied for got rejected, only 1 or 2 led to interviews, and the rest were either "under screening" or gave no response after I finished the assessment.

But I still open the spreadsheet every day to check the status of all positions that haven't rejected me one by one.

I am studying Chinese Language and Literature at Henan Normal University. I originally planned to work as a teacher after graduation, or take civil service and public institution exams. In the first semester of my junior year, I found that four out of five friends who were one year ahead of me were unemployed after graduation. All of them had prepared for civil service and public institution exams, but none of them passed. I suddenly realized that the path everyone is crowding into is not necessarily more stable than market-oriented employment.

So I immediately started looking for internships. My first internship was at BlueFocus, where I worked as a third-party service provider for L'Oreal, and that was where I first felt the impact of AI.

In the past, we had to contact dozens of small-sized KOCs to shoot product photos and make repeated revisions. Later, we only needed to put the product into an AI tool to automatically generate display images. The team in charge of UNIQLO next to us went even further: with 9 photos of the product provided by KOCs from the front, back and side views, AI can generate mirror selfies that look exactly like they were shot by real people. The supervisor asked at a meeting: "In the future, will it be enough to just buy the posting rights of accounts, and we don't even need real-person shooting?"

Later I joined Hello Inc. as an overseas market operation specialist, and now I am interning at a leading overseas-oriented company in Shanghai, using AI to generate advertising materials.

My leader said he would help me apply for a full-time conversion quota, and if there is no opening in our team, he would refer me to other teams. But I dare not miss the autumn recruitment. Many students online also share similar experiences: 5 people are recruited for summer internships, but there may only be 1 full-time spot in the end, or even none; the leader may genuinely want to keep you, or just hope you finish the work at hand first. So I keep applying for resumes while doing my internship.

I applied for Yutong, a local enterprise in Henan, and got screened out the next day after submitting my resume. I applied for 6 positions at Alibaba and got 4 rejections, and I didn't even dare to apply for ByteDance. My personal feeling is that leading enterprises in second- and third-tier cities often set stricter academic requirements, while leading internet companies will make trade-offs between academic background and practical experience. As long as your internship experience is sufficiently targeted, graduates from non-double-first-class universities can also advance to the next round of assessment.

At present, the only interview I have got from a leading internet company is for the AIGC Video Creative Production position at Pinduoduo.

The interviewer kept asking questions: What models have you used? Which one do you use most often? What is your workflow? How do you adjust the settings after generation fails? Where can the product be further improved? Adding more "AI" keywords to your resume may trick the automated screening system, but you will definitely expose your lack of relevant knowledge during the interview. Non-technical positions do not necessarily require you to understand the principles of models, but you must be able to present videos, images and commercial projects, and explain every step clearly. Certificates are not very useful. What really matters is your portfolio and internship experience.

Right now I can only be counted as someone who can use AI, not someone who understands AI. When my Agent program gets stuck, I will change the prompts and replace the materials, but I don't know what the problem is at the underlying level. I learned to use CapCut, AE, PR and generation tools all by myself outside of school. Preparing early does not guarantee success, but if you don't prepare, you may not even get the qualification to be considered.

02. Embodied intelligence positions are less competitive than algorithm roles, but they require highly relevant internship experience

Xiaobin ｜ Class of 2027 master's student majoring in Software Engineering at a top 985 university in Beijing

I am a master's student majoring in Software Engineering, now in my second year of graduate school and will graduate next year.

Actually, I started my autumn recruitment pretty early. From the early batch in July to now, I have applied for some positions at top tech companies based in Beijing, I have attended interviews at 3 or 4 companies so far, and I am waiting for their offer notifications.

I haven't applied for a huge number of positions, I want to take it step by step. The competition for the early recruitment batch is quite fierce, and some senior schoolmates also suggested that I apply later.

I mainly applied for embodied intelligence algorithm positions. I chose this direction because I had previous internship experience working on Agent products at a top tech company, so I am quite familiar with this field and can present relevant experience clearly in my resume. But after attending so many interviews, I increasingly feel that pure software development may not have much promising future. It's not that software development is not good, but that the competition is too cutthroat.

For this year's popular AI positions at top tech companies, such as large model algorithm, embodied intelligence algorithm, and AI Infra R&D, each of them is fought for by a large number of people with prestigious university backgrounds and internship experience at leading tech firms.

So now I am looking at positions in the embodied intelligence direction. There are also many Agent positions in this field, which are less competitive than large model positions and offer quite good salaries. But the problem is that I don't have relevant experience. The embodied intelligence field has extremely strict requirements for internship experience, you can hardly get in without a targeted internship, so I am hesitating now: it's too late to make up for relevant internships, the time is too short to add to my resume, I can only try to submit applications first.

I think the changes between this year's campus recruitment and last year are quite obvious. Last year, top tech companies already started recruiting AI positions on a large scale, but there are two biggest differences this year: First, the coding ability requirements for AI positions have become higher, the recruitment needs for many algorithm positions have changed accordingly, and the requirements for candidates have naturally risen. The ability to write original code with Python and C++ is no longer that important; Second, the total number of positions has decreased instead, and the demand is more refined. It seems that AI positions account for a high proportion, but compared to last year, many positions are no longer available.

I am not surprised at all by the high proportion of AI positions at top tech companies, I even think 70% or 80% is a low number. Under the current trend, what positions are completely free of AI? But it's worth noting that AI positions vary a lot, positions for people who use AI are also AI positions, and positions for people who develop AI are also AI positions, but these two types of positions are completely different in terms of recruitment standards.

From my observation, the gap between students who started preparing for AI positions very early and those who just started paying attention now is clearly visible in real recruitment scenarios, which is mainly reflected in internship and project experience. For example, students who have written project documents by themselves can handle many questions in interviews more smoothly.

As for the AI background requirement, I don't think you need to be too anxious. When top tech companies recruit for AI positions, they don't set very strict limits on majors, what they really care about is your academic qualification and internship experience. I know many senior schoolmates whose undergraduate majors have nothing to do with AI, but they still got AI product positions and perform very well. The key is to get into that environment as early as possible, even if you start with a marginal position.

Fresh graduates without AI background will not be eliminated by top tech company recruitment. It's true that recruitment at top tech companies has certain requirements for AI background in the past two years, but this industry is changing so fast. The opportunity you see now may have a higher threshold by the time you are ready, and with the current speed of AI development, the AI position you get this year may not even exist next year.

03. AI positions are everywhere at top tech companies, I switched from Java development to the AI application layer

Dangdang | Class of 2027 undergraduate majoring in Computer Science at a Shanghai university

I am an undergraduate majoring in Computer Science at a university in Shanghai. I have lost count of how many resumes I have submitted during the campus recruitment period. So far I have received 2 interview notifications, one is for the AI Product Assistant position at a top tech company, and the other is for an engineer position at an AI startup.

My original career plan was to become a traditional back-end development engineer, taking the business development path. But seeing the overwhelming number of AI positions in this year's campus recruitment at top tech companies, I had to re-evaluate my career direction. Now my adjusted plan is to first join the AI application layer or tool chain positions at a top tech company, and supplement my understanding of algorithms and models in practical work.

Regarding my feelings about AI campus recruitment at top tech companies, I do feel the huge impact: the proportion of AI positions is far higher than I expected, but top tech companies are still my first choice for job hunting, however my mindset is completely different from when I first entered university. In the past I thought working at a top tech company was just writing code, but now I think joining a top tech company is to get access to the most cutting-edge AI implementation scenarios, otherwise I will easily be left behind by the industry.

I also talked to several senior schoolmates who have already graduated. They think this recruitment wave comes after the large model technology has passed the simple "model training" stage and entered the "implementation and application" stage. A senior schoolmate working at a top tech company advised me, don't fixate on algorithm positions, pay more attention to how AI empowers existing businesses, because what top tech companies urgently need now are engineering talents who can get things done and deliver practical results.

To improve my competitiveness in AI job hunting, I did several specific things: first, I built a campus knowledge base Q&A system using large models; second, in my last internship, I tried to use Copilot to assist me in writing business code.

In the recruitment market, the gap between students who prepared early and those who just started paying attention now is very obvious. This gap is mainly reflected in the completeness of the technical stack and the depth of understanding of AI products. Students who prepared early can clearly explain the advantages and disadvantages of current large model architectures and have relatively mature GitHub projects; while students who paid attention to this trend late often only stay at the level of "having used AI", and lack systematic project support.

I think "being able to use AI" and "understanding AI" are completely different in job hunting. Being able to use AI means being proficient in using various tools to improve work efficiency, while understanding AI means grasping the principles of models and being able to optimize model structures. I didn't give up submitting applications just because I still have deficiencies in some underlying principles. On the contrary, my daily learning and accumulation in the AI application layer made me more daring to apply for those engineering implementation positions.

As for students without AI background, I think "universal AI literacy" has become a trend. One of my classmates majoring in marketing is now learning AI image generation and data analysis, because top tech companies require all positions to have AI literacy.

I spent a lot of time practicing algorithms and data structures at school, which are the foundation for understanding AI, but now that's not enough, I have to quickly build an AI knowledge framework on this foundation. All in all, facing the all-in AI strategy of top tech companies, I think I can barely keep up, but I have to run to not fall behind.

04. If you want to land an AI position at a top tech company, you must attach great importance to targeted internship experience

Xiaolu ｜ Class of 2026 master's student majoring in Psychology at a top university in eastern China

I am a graduate of the class of 2026, and I have just officially joined a top tech company to work in the product direction. I studied psychology for both my undergraduate and postgraduate degrees. The reason I got this offer in campus recruitment is that I interned there last year, delivered good results, was recognized by my supervisor, and naturally got the full-time offer in autumn recruitment.

Luck and accumulation both played a role in this process. I started doing internships since my sophomore year. At first I just wanted to find something to do during holidays, I did market research, user research, marketing, and later switched to the product direction. I have accumulated several internship experiences one after another at Xiaohongshu, ByteDance, Tencent and fast-moving consumer goods companies. I worked on commercialization at ByteDance, and also did a product internship related to AI at Tencent.

In August this year, the 2027 campus recruitment programs of various top tech companies were launched one after another. All top tech companies are going all in on AI, the changes are huge, when I applied