THESIS: Leveraging Machine Learning with LLMs and RAGs to Process Applicant CVs

High Level Description
Tools for automatically searching for optimal candidates with CVs often use keyword-based approaches. However recent advances in Large Language Models (LLM) and more specifically the Retrieval-Augmented Generation (RAG) architecture have a large potential for improving this search. These tools are valuable both for recruiting new employees as well as matching existing consultants to new missions. Candidate CVs can be converted embedded into a high dimensional vector-space and the text-based search query can also be embedded in this space to perform similarity operations. The RAG architecture can be used to improve this concept by combining the retrieval with an LLM to guide the search, report the top candidates and motivate why they could be a good fit compared to others.

Project Description
This project focuses on the development of a machine learning architecture using LLMs and RAGs to match prospective candidate CVs with text-based search queries. Key objectives include:

• Background: Perform a comprehensive study of existing ML recruiting tools and models to compare them and motivate which architecture is best suited to solve the problem.
• Implement models: Implement the pre-selected models using Python and finetune using cloud-based services.
• Testing and Validation: Test and compare the performance of state-of-the-art models in terms of accuracy, speed and cost.
• UI Creation: Design and build a user interface, allowing users to perform searches on a simulated database of CVs
  

The dataset used in the thesis will not use data from real people and instead use an open dataset of CVs, but the pipeline for real data should be the same. A manually labelled dataset with the best candidates for different roles will have to be created to enable comparison between the different models using suitable metrics.

Who are we looking for?
We are seeking a master's student with a background in Machine Learning, Data Science, Computer Science, or related fields to join our project. While previous knowledge of PyTorch or TensorFlow, cloud platforms and API based LLMs is beneficial, it is not required. The ideal candidate should have:

• Proficiency in Python and commonly used data science tools.
• An interest in Natural Language Processing (NLP), Information Retrieval, and applied AI.
• Motivation to gain valuable insight in the recruitment and consulting domain while becoming familiar with state-of-the-art AI tools.
• Willingness to build both backend systems and user-facing tools.
• Fluency in Swedish, both written and spoken

Purpose
The primary purpose of this thesis is to develop an automated AI system that can intelligently rank and match candidate CVs to job roles using LLMs and RAG techniques. The system will enable more accurate, efficient, and insightful candidate searches—both for recruiting new employees and assigning current consultants to new missions. By combining text embeddings, vector similarity, and LLM-powered explanations, the project aims to significantly enhance traditional keyword-based search approaches.

The thesis project can be published and used in your personal portfolio as well as in company marketing. Include Resumé/CV and cover letter in your application.

An Exciting Journey with Knightec Group
Semcon and Knightec have joined forces as Knightec Group. Together, we are Northern Europe's leading strategic partner in product and digital service development. With a unique combination of cross-functional expertise and a holistic business understanding, we help our clients realize their strategies – from idea to complete solution.

Practical Information
This is a Thesis position, located at our office in Solna, Rättarvägen 3. Start date

Please submit your application as soon as possible, but no later than If you have any questions, you are welcome to contact Myko Smid. Note that due to GDPR, we only accept applications through our careers page.