Client

Case study

Task

Develop an explainable competence archive that reconstructs professional competencies from real project work and compares them with competency requirements extracted from job descriptions.

Idea

Professional competencies are traditionally communicated through static documents such as resumes or job titles. In reality, however, competencies emerge continuously while solving real-world problems.
‍
As collaboration between humans and AI becomes an increasingly common part of professional work, these interactions create a new form of structured project evidence.
‍
This project explores how documented project work can be transformed into an explainable competence archive that makes professional competencies transparent, navigable, and comparable with the requirements of the job market.

Project Highlights

- AI pipeline for reconstructing competency profiles from real project work
- Analysis and standardization of more than 10,000 extracted tasks
- Extraction of tasks, methods, competencies, and supporting evidence
- Development of a hierarchical Competence Taxonomy
- Explainable AI job matching between competency profiles and job requirements
- Interactive Competence Archive for evidence-based competency exploration

Tech Used

- Python
- OpenAI API
- Large Language Models (LLMs)
- Embeddings
- Semantic Search
- Clustering
- Information Retrieval
- Explainable AI
- NumPy
- Pandas
- Scikit-learn
- JSONL & TSV Data Pipelines
- Visual Studio Code

Challenges

One of the biggest challenges was transforming highly unstructured project histories into reproducible and explainable competency profiles. This required deterministic AI workflows capable of separating tasks, methods, competencies, and other knowledge objects while minimizing semantic duplication across thousands of extracted entities. Another challenge was designing a hierarchical competence taxonomy that supports both reliable competency analysis and automated job matching. Finally, the resulting competency profiles had to be presented through an intuitive interface that makes professional competencies transparent and their supporting evidence easy to explore.

Current MVP

✔ AI Competency Reconstruction
✔ Competence Taxonomy
✔ Competence Archive
✔ Explainable Job Matching
✔ Interactive Evidence Navigation