CV inflation and traditional coding tests offer only a partial view of a candidate’s abilities, making it difficult for organisations to assess real-world competence and for candidates to be evaluated fairly on their true skills. South Korean scale-up Codepresso looks to change that.
Codepresso has been challenging traditional hiring tests in Europe since it opened offices in Luxembourg in 2024.
Co-founded in 2019, by IT expert Donghun Lee, Codepresso started life as a hands-on IT training company working with corporations and public sector clients. In 2022, they started to get requests from customers for skills assessments to support hiring and assess internal capabilities. Two years later, Codepresso launched its assessment platform for hiring, internal mobility and upskilling.
The firm’s relationship to Luxembourg followed a similar timeline. “Our connection with Luxembourg began in 2022, when we first visited the country through an invitation from Korea’s Ministry of Science and ICT,” explains Sebastian Lee, the Global Business Team Lead. Over the following two years, they returned several times.
Luxembourg stood out as a natural springboard for its business-friendly environment, its role as a strategic gateway to the European market, and its strong geographical and institutional connectivity across Europe.
“What ultimately solidified our decision was the level of trust and openness we experienced after establishing our entity. Despite being a foreign-founded company, we have consistently been treated as a Luxembourg-based company,” Lee says.
As evidence he explains that the firm joined the Luxembourg stand at VivaTech, Europe’s biggest tech event, in 2025, among other support.
Codepresso’s goal is clear: to offer more than traditional coding assessments and subjective interviews. “Traditional CVs offer useful signals about experience and self-reported technical skills, while conventional coding tests, often centred on algorithms and data structures, provide only a partial view of a candidate’s readiness for day-to-day work,” Lee says. Likewise, take-home assignments or internal project-based tests may more closely resemble real-world work. But they often demand considerable time and effort from internal engineering teams.
Lee says that many core aspects of actual job performance, including interpreting requirements, setting priorities, translating intent into working code, and maintaining code quality and performance, tend to fall outside the scope of these methods. And the gap becomes most visible in project environments that closely reflect real work conditions. “In such contexts, candidates with strong CVs and high test scores may still struggle with practical execution: misunderstanding requirements, opting for unnecessarily complex solutions, or relying on tools such as AI without sufficient contextual judgment.”
Lee says that the challenge lies in the structural limitations of traditional approaches to fully capture how individuals define problems, make decisions, and execute work in realistic settings.
Focus On Solution Quality
Codepresso’s end-to-end project-based assessments take 3-4 hours to complete, and are based on role-specific competency models and real industry requirements. The models are supported by a library of more than 2,000 assessment items, enabling both realism and scalability in skills evaluation.
“Assessment results are mapped to clearly defined proficiency levels: Entry, Beginner, Intermediate, and Professional,” says Lee, adding: “In parallel, candidates are evaluated across specific skill dimensions defined within each competency model, with skill acquisition levels classified as Excellent, Needs Improvement, or Insufficient.”
Real-world Impact
The solution is already having an impact on hiring decisions. Among its clients, Codepresso counts a Korean startup that froze hires rather than find an in-house testing solution. The result: hiring resumed and decisions were based on evidence rather than intuition.
Beyond hiring, the tool is used by employers as an organisational asset. Hyundai Mobis used it to assess approximately 3,000 participants as part of a mobility talent development initiative, with around five percent ultimately selected through a validated assessment process. “Those selected then progress through ongoing assessment and structured upskilling and reskilling programs, with outcomes including hiring or internal role transitions,” says Lee.
In another example, Korean Air’s software developers participate in annual, role-based competency assessments through Codepresso, the results of which are used for making decisions related to compensation, promotion and long-term workforce planning.
“This enables an organisation to manage technical skills at an organisational level using structured, objective, and quantitative skill data,” says Lee.
Measuring AI Literacy
As AI shows promise at writing code itself, in the future the ability of programmers to write code efficiently will be less important than the ability to define problems and exercise sound judgement on how to solve them. Lee reckons that across most professions, the gap between AI adopters and those not using AI will widen. In short: the candidates who get hired will be those who can effectively leverage AI in their work.
That is why in 2026 Codepresso is launching AI Fluent, an AI proficiency and application skills assessment. Intended for a wide range of roles, including developers, sales, marketing, HR, finance, and accounting, it focuses on how individuals apply AI in real-world scenarios and technical proficiency for execution capability to implementing, operating, and applying AI tools.
Throughout the process, Lee stresses that HR and human judgement remain critical in the hiring process, with skills-based hiring serving as a complement, not a replacement.
