Amid the repetitive cycle of resumes, interviews, and assignments, this service emerged as a pivotal turning point in developer recruitment.

Have you ever invested time and resources into the hiring process, only to discover that a candidate’s real-world skills fall short of expectations? Identifying genuine talent can be challenging—especially when résumés are refined, interviews are well-rehearsed, and assignment submissions appear flawless. To navigate this complexity, Archisketch leverages SkillCertify, a robust developer assessment solution, to deliver transparent, cheat-resistant evaluations that reveal candidates’ true technical capabilities.

Q. Hello! Could you introduce yourself and your company?

Hello, I’m Da-jeong Yeom, the HR Manager at Archisketch, where I oversee all aspects of human resources—from recruitment and performance evaluations to fostering a positive organizational culture.

Archisketch is a leading IT company specializing in 3D interior design solutions. Guided by our mission, “to make interior design easy for everyone” we are committed to transforming a traditionally opaque industry. Our goal is to enhance transparency and trust in the interior design market through cutting-edge technology that empowers users and simplifies the design process.

Limitations of the Traditional Developer Hiring Process: Heavy Burden on Development Teams and Low Accuracy

Q. Before using Codepresso’s service, your team designed its own hiring assignments. Can you walk us through how that process worked?

Our process typically began with résumé screening, followed by interviews. Only candidates who passed these stages were given developer-created assignments. In other words, the flow was: [résumé → interview → assignment].

Q. What challenges did you encounter with that approach?

That process consumed a lot of resources from our developers. They had to review résumés, attend interviews, send and manage assignments, respond to candidate questions, and so on. Over time, this led to growing fatigue within the team. The effort required to hire even a single person often felt disproportionate to the uncertain outcomes. At first, we held interviews before giving assignments. But since interviews demanded so much time and energy, we realized a shift was needed. By introducing skill evaluations before interviews, we could filter candidates earlier and reserve our developers’ time for those who had already demonstrated strong technical ability. Partnering with Codepresso made this change seamless, significantly easing the hiring workload on our team.

Q. After changing the hiring process, did you notice any difference in assignment quality or candidate response rates?

Previously, most candidates moved on to assignments after interviews. With the new process, however, we saw a noticeable drop-off during the evaluation stage. While this initially raised the question of whether we should adjust the order again, we ultimately decided to maintain the evaluation’s difficulty at a fair but firm level. Our reasoning was clear: if applicants were unable to demonstrate the necessary foundational skills at this stage, they were likely not aligned with our expectations.

Q. When you created assignments in-house, candidates had the chance to research. With proctored evaluations, did you notice more candidates giving up midway?

Yes, that became a clear issue. For example, we once had a candidate who successfully completed an assignment and passed the hiring process, only for us to discover later that they struggled with tasks that should have been simple for someone at that level. It turned out the assignment results hadn’t truly reflected their own skills.

With today’s advanced AI tools, even someone without a development background can generate working code through online searches or AI assistants. That experience highlighted for us the need for a more reliable and integrity-focused evaluation service — one that could ensure the results genuinely represent a candidate’s capabilities.

Q. After introducing evaluations, did the number of candidates progressing to interviews decrease significantly?

Yes — in fact, the drop was greater than we had anticipated. Only a small number of candidates advanced to the interview stage. At one point, we even decided to interview applicants who had scored below our target threshold, but the outcome was always the same: “they’re still not at the level we need.”

From that experience, we established a clear standard. Today, we only consider candidates scoring in the high 70 score as interview-worthy, while those in the 80 score progress directly to interviews. In other words, interview decisions are now driven by evaluation scores, giving us a far more reliable filter for talent.

Q. Did this make you think about increasing the number of applicants?

We considered it, but our approach has shifted toward evaluating each candidate with far greater care than before. Our guiding principle is now: “Even if we hire only one, let’s make sure it’s the right person.”

While evaluation costs can feel like a burden, we see them as an investment in avoiding much higher opportunity costs. Hiring the wrong candidate without proper verification can lead to wasted time and resources. Onboarding and training require contributions from multiple employees. If the new hire leaves after just a few days or proves to be a poor fit, the cycle of re-hiring creates talent shortages and substantial losses.

Taking all of this into account, we believe our current evaluation-first approach is not only reasonable but also the most effective way to build a stronger, more sustainable team.

Background and Actual Use of Codepresso

Q. What was the main reason or background for adopting Codepresso’s service?

The decision stemmed from two major challenges: the significant recruitment burden placed on our developers and the difficulty of truly assessing candidates’ skills. With the rise of AI tools making it easier for applicants to rely on external help during assignments, concerns about fairness and accuracy also grew. We realized we needed a more objective, transparent, and reliable online coding test solution — and that’s when we turned to Codepresso.

Q. After adoption, which tracks or functions are you mainly using?

We leverage a variety of tracks tailored to the specific skill sets required for each role.

For Software Backend Engineers, we focus on Java, Spring Boot, and Kotlin.

For AI Engineers, we use HuggingFace, while for AI Backend Engineers, we apply LangChain.

When it comes to Technical Project Managers (TPMs), we assess their knowledge in both Software Engineering fundamentals and Prompt Engineering.

Looking ahead, we plan to extend this approach to Frontend Engineers as well.

Q. How do you actually use the “competency reports” provided by Codepresso in your decision-making?

The reports are really detailed, which makes them valuable not only for developers but even for non-developers like myself. They clearly explain why a certain score was given and break down results into sub-scores, which significantly increases their reliability.

For example, we recently had a candidate who scored very high. However, the report flagged that they had copy-pasted examples. And indeed, upon review, this turned out to be true — so we immediately disqualified the applicant. In reality, neither I nor our developers have the time to thoroughly check every detail. Without that specific comment in the report, I might have assumed the candidate was a strong fit and advanced them to the interview stage. That single insight saved us from making the wrong decision, which shows just how critical these reports are in our hiring process.

Q. While using the service, was there any part you wished could be improved?

At one point, I needed the ability to duplicate evaluations — and that feature was recently added, which solved the issue. I also had a request to separate questions within a set. Overall, though, most of the needs I’ve raised have already been addressed, so the service has continued to evolve in ways that meet our requirements.

Positive Changes and Effects at Archisketch After Adopting Codepresso

Q. What was the biggest change you noticed after introducing the skill evaluation service?

The most immediate impact was a significant reduction in our developers’ workload. Previously, every interview interrupted their workflow, making it difficult to stay focused on projects. Now, with evaluations handling much of the screening process, our developers can concentrate almost entirely on their core work. This improvement in focus not only benefits individual productivity but also drives broader growth for the company.

Q. With developers’ workload reduced, did you notice any changes in the quality of candidates you interview?

Yes, absolutely. For instance, there was one candidate we were eager to hire, but negotiations ultimately didn’t work out. Because we had evaluated them so highly, it was clear they were a top talent with multiple offers on the table. This shows that we are now meeting candidates we truly don’t want to lose.

Previously, when our team interviewed more than 10 candidates, fatigue often lowered the quality of each interview — which could even have a negative effect on our hiring brand. Now, with fewer but stronger candidates, the quality of interviews has improved significantly, and this has become another positive outcome of adopting the evaluation-first approach.

Archisketch’s Hiring Vision and Future Plans

Q. What direction or type of talent is Archisketch pursuing in future recruitment?

These days, instead of strictly separating frontend and backend roles, there’s a growing preference for full-stack engineers who can adapt across different areas. More importantly, however, we’re looking for people who can align with and adapt to our company culture.

During interviews, we place strong emphasis not only on technical skills but also on cultural fit. As a startup, we experience frequent changes — business directions shift, teams evolve, and roles and responsibilities are redefined. We value people who can embrace this environment and say, “Okay, I’ll take on this task too!” with flexibility and positivity.

We also maintain a flat culture where all employees, including the CEO, use English names. Candidates who can easily adapt to this open and collaborative atmosphere are especially welcomed at Archisketch.

Q. Lastly, do you have any recommendations for other HR managers who are facing similar difficulties with developer recruitment?

In large companies, HR teams often include staff with technical backgrounds who can dedicate significant resources to developer hiring. But for most SMEs, this simply isn’t realistic. Developers are expected to both build products and participate heavily in recruitment — a combination that is overly demanding.

For companies in this situation, adopting Codepresso can make a real difference. It not only eases developer fatigue but also empowers HR managers to better support their teams, creating a more balanced and effective hiring process.