Our work / Škoda AI career assistant

Introducing a fully-equipped career with an AI assistant under the bonnet at Škoda Auto a.s.

Web presentation

Requirements

ŠKODA AUTO is one of the largest employers in the Czech Republic with a wide range of open positions across manufacturing, technical development, IT, and other fields. Candidates on the career portal often struggled to navigate the extensive list of vacancies, making it difficult to quickly identify which roles best matched their qualifications and experience. They had to manually browse through dozens of open positions and assess themselves whether they met the requirements. This process was time-consuming and could lead to overlooking relevant positions. The existing robust system also often discouraged candidates from advanced searching. 

In collaboration with Škoda IT, we developed an AI career assistant that allows candidates to upload their CV directly on the website. The system extracts text from the CV, classifies candidate skills, and automatically matches them with currently open positions. The result is a personalised list of up to 3 best-matched roles including profile match scores (Profile Matching 0–100%). 

Candidates receive immediate and relevant feedback. Simply upload the CV and within seconds get tailored recommendations. This increases conversion on the career portal, reduces the risk of ideal applicants getting lost among resumes, and improves the overall user experience. 

Two-stage AI pipeline – classification and intelligent recommendation

In designing the solution, we chose an innovative two-stage CV processing approach using Azure OpenAI: 

  1. CV classification: The uploaded file is first processed by Azure Document Intelligence, which extracts text even from scanned documents and images. Subsequently, Azure OpenAI classifies the candidate's skills, education and experience into predefined categories (fields, job categories, locations). The system also assesses whether the document is a genuine CV (ValidCvScore 0–10). 
  2. Pre-selection of relevant positions: Based on classification results, the system automatically filters open positions that best match the candidate’s profile. The AI model then works only with this narrower group of positions, ensuring more accurate and faster recommendations. 
  3. Position recommendations: In the next stage, the Azure OpenAI model compares detailed descriptions of the filtered positions with the candidate’s profile and selects up to 3 best matches. It calculates a Profile Matching score for each position based on four dimensions: skills match, relevant experience, education, and additional factors.
View the work result for desktop and mobile devices

Delivered Services

Technology

Web presentation
It is important for us to have a direct contact with our clients. We can consult with you both in-person and online. Please do not hesitate to contact us with your inquiry or further specification of your request.
Petr Lebeda
Petr Lebeda Sales & Consulting Manager +420 723 484 557

Do you want to know more? Ask us a question.