As the valuation of the AI programming track rose and competition intensified, competition for talent from head starters also escalated. Business Insider reports that companies such as Cursor, Cognition, Reprit, Base44 are downplaying traditional curriculum vitae screening and standardized interviews and moving to a more direct search for engineers who can deliver products quickly.
The founders went out and dug.
These companies rely less and less on public delivery and are reaching out to targeted candidates. According to the report, during the rapid expansion phase of Cursor, CEO Michael Truell will personally look for X and GitHub information, looking for engineers, former entrepreneurs and employees from Notion, Figma and others.
Exposure patterns are also more intensive. The former employee stated that the candidate might have received an email from the founder before being invited to coffee and to the office. The Cognition, Emily Cohen, who is in charge of manpower, also indicated that she had travelled overseas to meet with candidates and even personally sent them to the airport for final communication.
Decline in CV weights
In such companies, the importance of resumes and LinkedIn materials is declining. The Replit Manpower Manager Stacey La Torre states that X has become one of the main sources of recruitment for companies and that people who come in contact with social platforms, meetings or other venues are shared within the team through special channels.
Base44 CEO Shlomo also indicated that the company valued the products and public works already produced by the candidates more than academic or curriculum vitae packaging. Actual past outputs are more persuasive than the standardized context for such teams, which emphasize rapid build capacity.
One day job instead of one day interview.
Traditional multi-cycle interviews and algorithms are being replaced by longer field missions. Reports indicate that companies such as Cursor and Kilo will arrange for more than one day of testing or a week of training to allow candidates to be directly involved in work close to their official positions.
In the case of Cursor, candidates will use the company ' s computer to carry out their tasks on a frozen version of the code library and will end up showing results. It was also mentioned that there had been a management candidate who had worked for a month, had been in contact with almost the entire team and had eventually not been recruited, reflecting the high threshold of such companies for new members.
According to Kilo CEO Scott Breitenother, this means sifting out those who do good paper, but do not adapt to high-paced entrepreneurship. According to him, some candidates would take leave to join the company ' s centralized workweek and, if adopted, to leave the company and join the team.
Interviews begin to examine AI usage habits
In addition to the trial work, the question of the interview itself is changing. The Chief Operating Officer of Rocket, Deepak Dhanak, stated that the technical interview would first ask candidates how much tokens are spent on work and personal projects each week to judge the frequency of use and the intensity of experiments on large model tools.
There is a general desire for greater autonomy and “founder mentality” among the employees. For example, Kilo asked candidates what different decisions they would make if they were the CEO of their previous job, and Base44 asked candidates about the targets for the coming years to judge whether they were adapted to a high-input, high-intensity entrepreneurial environment.
Allow the use of AI tools in interviews
Another notable change is that some companies no longer see the “prohibition of AI” as a reasonable precondition for interviews. Cognition believes that interviews should also allow candidates to use AI tools if the post itself is required to perform the development task.
Base44 and Reprit also take a similar approach. It was mentioned that some companies would require candidates to complete the project directly with the interviewer during the interview or to observe their ability to collaborate with AI, dismantle problems and deliver quickly by providing on-site information on how to construct a product.
Overall, AI's recruitment criteria for programming start-ups are moving from “can you answer the questions” to “can you make a product right away”. As the competition for talent continues to warm, such recruitment processes, which place more emphasis on combat, speed and primary AI, may spread among more AI companies.