Lidija Jovanovska
Data Scientist | NLP & Machine Learning Specialist
About Me
I’m a data scientist specialising in building and deploying ML solutions. Since publishing my first research paper in 2019, I’ve been driven by a curiosity to solve complex problems and push the boundaries of data science and AI.
Currently, I’m part of the innovative team at Sportradar, where I focus on building NLP-driven solutions for entity mapping and semantic search. One of my proudest achievements has been deploying a semantic search engine that significantly improved mapping accuracy for various company services. I also get to have a little fun building GenAI products for sports commentary. Prior to this, I honed my skills at a real estate startup, developing machine learning models that boosted price estimation accuracy and strengthened client relationships.
I thrive on the intersection of innovation and impact. Whether it’s deploying machine learning models into production or fine-tuning language models for semantic annotation, I believe in the power of data to transform industries. I’m always eager to learn, explore new techniques, and bring creative solutions to the table. And if you ever want to chat AI, NLP, or how to make machines understand us a little better, feel free to reach out!
Skills
Python
- Extensive experience with libraries like Pandas, Numpy, Scikit-learn, Pytorch, Plotly, Huggingface libraries
SQL
- Proficient in writing complex queries and working with relational databases
FastAPI, Django
- Web development experience building robust APIs
Docker, Kubernetes
- Containerization and orchestration of ML models in production environments
Natural Language Processing (NLP)
- Named Entity Recognition (NER), Semantic Search, Text Classification, LLMs
Machine Learning
- Developing and deploying ML models for tasks such as entity mapping and price estimation
Data Visualization
- Using Plotly, Matplotlib for creating insightful visualizations to communicate findings
AWS
- Experience with services such as Redshift, S3, Sagemaker for model training and data handling
MLFlow, Kubeflow
- Managing machine learning lifecycles, model versioning, and experiment tracking
Git + CI/CD Pipelines
Version control, continuous integration and continuous deployment practices
Kafka
Real-time data processing
Work Experience
Sep 2022 - Present · Ljubljana, Slovenia
Key responsibilities
- Develop and implement NLP solutions for entity mapping and semantic search
- Collaborate with cross-functional teams to integrate ML models into production systems
- Build GenAI products for sport (e.g., match commentary)
Technologies
- Python, Kubernetes, FastAPI, Docker, SQL, Kafka, AWS (Redshift, S3, Sagemaker), MLFlow, Kubeflow, Git
- Huggingface, SBert, Pytorch, Pandas, Numpy, Plotly
Achievements
- Successfully deployed a semantic search engine boasting 98% mapping accuracy
- Developed a match commentary solution that will be integrated across different services
Sep 2021 - Apr 2022 · Ljubljana, Slovenia
Key responsibilities
- Developed ML models for real estate price estimation
- Created and maintained model architecture documentation for external validation processes
- Built a package for predicting energy efficiency of real estate
Technologies:
- Python, SQL, Git
- Django, Pandas, Numpy, Scikit-learn, Plotly
Achievements
- Improved price estimation accuracy by 10%, leading to increased client satisfaction
- Contributed to the data and model documentation which helped lock key clients
Oct 2019 - Sep 2021 · Ljubljana, Slovenia
Research Focus:
- Developing an ontology for AI entities and processes
- Using semantic technologies to annotate a data corpus (papers, algorithm documentation)
- Fine-tuning Named entity recognition (NER) models for automatic semantic annotation
Key Contributions:
- Developed tools for semantic annotation of AI algorithms
- Collaborated on multiple research projects within the Knowledge Technologies department
- Co-authored 2 research papers on semantic web technologies
Skills Developed:
- Advanced machine learning techniques
- Semantic web technologies (RDF, OWL, SPARQL)
- Research methodology and scientific writing
- Collaboration in an academic research environment
Education
International Postgraduate School Jožef Stefan
- Oct 2019 - Oct 2021
- Took courses in Machine Learning, Natural Language Processing, Semantic Web
- Masters Thesis: Semantic annotation of Machine Learning and Data Mining Algorithms
- Member of the IPS Student Council and Principal organiser of the 2021 IPS Student Conference
Faculty of Computer Science and Engineering - Ss. Cyril and Methodius University in Skopje
- Sep 2015 - Sep 2019
- Took courses in Discrete Mathematics, Calculus, Probability and Statistics, Computer Vision, Robotics, Network Science, Natural Language Processing, Computer Graphics, Bioinformatics
- Laboratory assistant for the course Computer Animation
- Published two research papers
Awards and Conferences
Best Documentation | HAMR Hackathon - ISMIR 2020
- Analysis of Chord Progression Networks using data from Ultimate Guitar
- Demo for generating chord progressions using graph traversal methods
Best student paper (3rd place) | CIIT 2019
- Awarded for the paper The Geographic Flow Of Music On Spotify
Presenter | Data Fair 2025
Attendee | International Conference on Machine Learning (ICML) 2024
Presenter | Data Science Conference 2023
- Deep-dive into the NLP solutions for entity mapping and semantic search at Sportradar
Presenter | MIPRO 2021
Presenter | DRMN+15 2020
Presenter | SiKDD 2020
Presenter | CompleNet 2020
Presenter | CIIT 2019
Presenter | CIIT 2018
Other Experiences
International Society for Music Information Retrieval (ISMIR)
Oct 2020
- Assisted with online conference management
- Moderated discussion panels
- Gained experience in virtual event coordination
Key Skills: Event Management, Online Moderation, Music Information Retrieval
Faculty of Computer Science and Engineering
Jan 2018 - Jun 2018
- Created comprehensive video tutorials for the Computer Animation course
- Assisted students during laboratory exercises
- Managed a Youtube channel with educational content
- Achieved Certified Autodesk Student Expert status
Key Skills: 3D Animation, Teaching, Content Creation, Autodesk Software
Neotel
Apr 2018 - Jun 2018
- Designed and manufactured drone parts using 3D printing
- Assembled a functional drone
- Experimented with photogrammetry software for 3D object generation
Key Skills: 3D Printing, Drone Assembly, Photogrammetry
Museum of the Macedonian Struggle Feb 2017 - Jun 2017
- Created an interactive 3D virtual version of a museum exhibition room
- Utilized Autodesk Maya, Unity, and WebGL for development
- Published a research paper on the learning environment creation process
- Contributed to innovative educational tools combining historical accuracy with user engagement
Key Skills: 3D Modeling, Game Development, Educational Technology, Research
Certifications
Note
These certifications demonstrate my commitment to continuous learning and expertise in key areas of data science and machine learning.
- Natural Language Processing Specialization | Coursera, January 2023
- Comprehensive training in NLP techniques, from basic text processing to advanced deep learning models for sequence modeling.
- Machine Learning in Production | Coursera, November 2022
- Advanced course on designing, deploying, and managing ML systems in production environments.
- Deep Learning Specialization | Coursera, July 2020
- In-depth study of neural networks, covering everything from basics to advanced architectures like CNNs and RNNs.
- Cambridge English: Advanced (CAE) | Cambridge, June 2014
- Certification of C1-level English proficiency, demonstrating advanced communication skills in professional and academic contexts.
Tip
For more information and to see my latest projects, please visit my GitHub profile
Want to get in touch? Contact me