When it comes to artificial intelligence, many think of the annual wage of millions, thesis of the conference and the school master of tsingbei. But today, more and more young people in “non-scientific classes” are entering this track. They learn statistics, computers, some of them not even ai. Without a prominent academic background, they have clear planning and solid experience, and they want to prove with their own experience that in the fast-temporal industry of ai, power works better than labels。
Reporters, pleasures, interns, lukage
Zenan, wong lilong, luber state
There's a man who's going to make a move
There's a man who's been hit by a mistake
"i can't think at first." imu, who studied statistics, was overstretched by the online “top paper anxiety” and even planned to postpone graduation. Until chatgpt was on fire, she saw a new way out: data modelling. At the post-graduate level, she did not dwell on her professional background, but rather looked at the job needs of “tailored” internships, which consisted of data analysis, quantification and, ultimately, modelling. Each experience is clear and clear enough to withstand the interviewer's digging, “the curriculum vitae is a door-knocking brick, each detail of which has to withstand a push”。
Compared to lin's steps, learning the structure of the computer system is “unexpected”. Zhang zheng only wanted to go home and get a job, and then went to the bottom of a factory. He went through three rounds of technical interviews all the way through the national processor project, which he studied dead. After only two months of entry on duty, he met with a transformation of the sector to ai, following which he made a large model of the national production calculator, “in white, catch up with the hard-headed”。
In many people's imagination, the ai industry's academic paper is impressive. But it seems to these cross-borderers that power is the real knock-on。
The head of a major factory's recruitment service revealed that in recent years there had been a shortage of technical, engineering and ecological staff. The job offers are not 985, 211, “both houses have opportunities. We look at hard power, suitability, enthusiasm, education as a reference”。
During the interview, three rounds of technical interviews surrounded his curriculum vitae. From the core logic to the risc-v basic operation, every detail is taken over, "there's nothing to be afraid of as long as you do it. Be confident, don't be shy”. In six months of his career, he felt the technology was too fast. The big model's thesis, the program, went on and on, "you won't catch up next week without learning." the same thing happened to lin-woo: "in the ai business, stopping to study is an automatic exit."
Not always
But there's a lot of stress
Outsiders look at ai practitioners and think that either they're “996 monks” or they're lucky to be able to hit the code. But in hayashi's eyes, real ai work is another kind of "tired."。
She does natural language processing in zong, day-to-day looking at models, adjusting parameters, optimising data and repeating them. A model runs for six months to see the effect, and it's too urgent. Outsiders seem to walk in the hallways and stare at computers, thinking they're touching fish, but that's when they're thinking, and they're turning their brains. “as opposed to physical work, intellectual work is not always done, but the spirit is always tense.” as long as there are technical problems left unresolved, lin said, he's going home from work, “that kind of psychological burden that no outsider can see”。
Zhang zhang's pace of work was two stages: no kpi had just arrived, and he was once worried about “not learning anything”. The sector was then transformed, with a sharp increase in intensity, and returned to the status quo of studies, “sometimes working overtime, but able to carry it and become used to it”。
Both mentioned that the most important aspect of ai's work was its ability to solve the problem, which was impossible to think about at once, but with a clear line of thought. Zhang zhang said, “it is only efficient to communicate, to ask, to be willing to share and to network knowledge.”
Cross-professional, shortboard must have. In narino's case, she studied statistics, and after her entry she clearly felt tired about the system architecture, the gpu operation and the model deployment. The base of zhang xian computer network is thin, and it works with occasional carding。
"what's wrong?" lin-yu relies on the company's tolerance period, "three to six months, allowing you to slow down, allowing me to adapt slowly to the pace of my work, and asking if you don't understand, this painful period will soon pass."
Zhang zhang works during the day and sees the technical course at night. In the case of network problems, the resolution process is taken down step by step, and notes are made to facilitate the recurrence。
“corporate culture is also strong.” according to lin, the technological platform, programming language, business scenes of different companies are different, and even with internships, entry is almost zero. I've been helped by the willingness of leaders and colleagues to teach.”。
Today, lin-woo has been able to take responsibility independently for model training and deployment, and zhang has been able to take on one side and take on the core of the calculus. Two once “go-and-goers” have taken their feet in the ai field。
I'm the one who needs the most
Applied talent
Who is the most lacking in the ai industry in the future? According to many practitioners, the most critical is applied talent。
"the base of the big model is now very mature, and deepseek is very capable." a senior algorithm engineer said, “how hard it is to land, how to integrate quietly into life and work, and how to solve real problems. That's the least of them."
For students who want to enter the ai field, the respondents suggested that the core base should be firmly established, knowledge of programming, mathematics and machine learning should be solidized, common skills such as linux, docker and computer networks should be acquired in advance, practical projects should be developed, skills should be kept up to date, and the ability to collaborate across jobs should be enhanced。
From “never think” to “stable heels”, the biggest feeling of the forest is that the ai industry is always open to people on the ground。
