Data Science, Recommender Systems
Aktualisiert am 22.10.2025
Profil
Freiberufler / Selbstständiger
Remote-Arbeit
Verfügbar ab: 22.10.2025
Verfügbar zu: 100%
davon vor Ort: 0%
Data Science
Recommender Systems
Algorithms
Bosnian
native
English
fluent
German
fluent (C1/C2)

Einsatzorte

Einsatzorte

Deutschland, Schweiz, Österreich
möglich

Projekte

Projekte

8 Monate
2024-11 - 2025-06

Agentic AI for a DeepResearch project

To create a multi-agentic system, supported by a Knowledge Graph, that automates the process of drafting a research paper. The system used multiple experts (OpenAI models) that ?collaborated? during the process of document drafting. The whole process was supported by a Knowledge Graph out of which we extracted useful information. Technologies used for this framework: LangChain, LangGraph, Smolagents, LlamaIndex, dspy. The project involved the use of Terraform and GitHub Actions (CI/CD Pipeline), via AWS. The initial application was deployed as a Streamlit app.
6 Monate
2024-05 - 2024-10

Open source project - PulseSpotter

PulseSpotter is designed to assist journalists in the task of identifying newsworthy topics that are likely to become popular. By gathering information from various news sources and analyzing patterns over time, the system suggests emerging trends by using AI and machine learning, saving journalists time on research and deciding which stories to focus on.
MediaLab Bayern
8 Monate
2024-02 - 2024-09

RAG (LLM)

Objective: an LLM chatbot to help with HR-related inquiries.
  • Approach: I led the development of an advanced Retrieval-Augmented Generation (RAG) system aimed at improving HR data retrieval processes. This system utilizes the Milvus Vector Similarity Search Database and OpenAI?s API to efficiently source and integrate extensive HR-related data, enabling it to respond to a broad spectrum of HR inquiries. We have also attempted to fine-tune an LLM for this task and the technologies used were PEFT (LoRA), and sophisticated Prompt Engineering.
1 Jahr 1 Monat
2023-09 - 2024-09

Spectral Image Object Detection

Objective: given a spectral image that depicts the value of the sensors?s readings, classify the signals and identify novelties (anomalies).
  • Approach: Given that there was not sufficient labeled data, I had to rely on selfsupervised machine learning paradigms. To satisfy the customer?s request for fast processing, I utilized one of the YOLO architectures. The system was developed with the mmyolo framework.
5 Monate
2023-08 - 2023-12

User Grouping in a Social Network

  • Objective: To foster communication among like-minded users through dynamic group formation based on their responses to specific questions.
  • Approach: Employed vectorization and K-Nearest Neighbors (KNN) for initial grouping, with a recommendation system for refined group alignment.
5 Monate
2021-04 - 2021-08

Fraud Detection for a Buy-Now Pay-later Platform

  • Objective: Fraud detection.
  • Approach: I designed and developed a fraud detection model for a Middle-Eastern Buy-Now Pay-Later platform. A critical step for this client involved data preprocessing, during which I employed a graph-based approach to identify cliques of fraudsters. This was the first successful machine learning project for this client.
8 Monate
2020-09 - 2021-04

Next-Best Finance-Related Action

  • Objective: To advise individuals on financial actions that enhance their chances of achieving specific goals.
  • Context: For an Asian bank aiming to offer actionable advice for long-term financial planning, such as securing a home purchase in 20 years.
  • Approach: Predictive machine learning models were developed to suggest optimal actions (example: obtaining a salary increase of at least 10%).
4 Monate
2020-05 - 2020-08

Recommender for Online Betting

  • Objective: To surpass the performance of Amazon Personalize?s Hierarchical Recurrent Neural Network-based recommendation system.
  • Approach: Adopted a recommender model based on Factorization Machines, significantly improving recommendations through targeted feature engineering.
8 Monate
2019-09 - 2020-04

Debt Collection

  • Objective: To predict the likelihood of loan repayment by bank customers, aiding in their segmentation for tailored communication strategies (email, SMS, or phone calls)
  • Achievement: Created a behavioral scoring machine learning model, now incorporated into Receeve?s collection approach.
8 Monate
2019-01 - 2019-08

Sensor Anomaly Detection

  • Objective: To detect unusual behavior in a sensor-monitored network.
  • Method: Established a baseline machine learning model predicting expected sensor readings, using deviations from this model to flag anomalies. Enhanced detection accuracy with a Graph Neural Network, analyzing the connectivity between sensors and observable areas.
9 Monate
2018-04 - 2018-12

Pricing Engine for a US Dispatcher

  • Objective: To predict vehicle transportation prices, providing dispatchers with a reliable basis for pricing.
  • Achievements: Enhanced the machine learning model?s accuracy by approximately 40% through advanced feature engineering and implementing a two-tiered regression strategy, combining residual and standard regression techniques powered by XGBoost.
3 Monate
2018-01 - 2018-03

Investment Prediction

  • Objective: To forecast investors? forthcoming decisions on company investments.
  • Approach: Transformed the challenge into a graph-based problem, treating investors and companies as nodes. Developed a Graph Neural Network to predict potential investor-company connections, integrating this model into StockFink?s prediction suite.

Aus- und Weiterbildung

Aus- und Weiterbildung

3 Jahre 2 Monate
2009-11 - 2012-12

Study - Computer Science

Ph.D., The University of Konstanz (Germany)
Ph.D.
The University of Konstanz (Germany)

  • Scholarship by the German Research Foundation (DFG)
  • Work Group: Algorithmics
  • Thesis: on request

3 Jahre 9 Monate
2005-10 - 2009-06

Study - Computer Science

Bachelor of Science, The University Sarajevo School of Science and Technology, Sarajevo, BiH, The University of Buckingham, Buckingham, UK
Bachelor of Science
The University Sarajevo School of Science and Technology, Sarajevo, BiH, The University of Buckingham, Buckingham, UK

  • Major in Computer Science, minor in Electrical and Electronic Engineering

Kompetenzen

Kompetenzen

Top-Skills

Data Science Recommender Systems Algorithms

Aufgabenbereiche

  • LLM
  • RAG (Graph-RAG)
  • Agentic AI
  • Recommender Systems
  • Deep Learning
  • Graph Neural Networks
  • Pricing Engines
  • Machine Learning
  • Dimension Reduction
  • Graph Embedding and Information Visualization
  • Algorithm Engineering (large data processing)
  • Numerical Optimization

Produkte / Standards / Erfahrungen / Methoden

Profile
The candidate received his doctorate in computer science from the University of Konstanz at the age of 25. During his academic career, he worked in the areas of dimensionality reduction and graph embedding, and his work has been recognized by the scientific community. As a (senior) data scientist, the candidate focuses on LLMs/RAGs, recommender systems, knowledge graphs, and classical machine learning. His most notable work concerns the development and implementation of a recommender system for the ARD audio library. He is also a fiction author.

Computer Skills
Python
  • Numpy
  • Pandas
  • scikit-learn
  • TensorFlow
  • Keras
  • PyTorch
  • PyG
  • Networkx
  • Matplotlib/ Seaborn
  • XGBoost/ LightGBM/ CatBoost
  • SpaCy/ NLTK

Other
  • AWS: 
    • deployment
    • Lambda
    • EC2
    • ECS
  • Azure: 
    • Databricks
  • Agile Development: 
    • Scrum
    • Kanban
  • Kubernetes
  • Streamlit
  • Spark

Scripting
  • Shell
  • Batch

Typesetting
  • LaTeX

Recommender Systems
  • collaborative filtering (linear models, matrix-factorization)
  • ensemble models
  • bandit-based models
  • graph-based recommenders
  • neural-network recommenders


Software

Falcone ? a graph profiling software based on multidimensional scaling techniques, written in Java


Deep Learning Frameworks

  • TensorFlow & Keras: Extensive experience in building and deploying neural network models.
  • PyTorch: Proficient in model development and experimentation, with a focus on dynamic computation graphs.
  • PyTorch Geometric (PyG): Skilled in implementing graph neural networks for complex relational data.
  • PyTorch Lightning: Familiar with this framework for scalable and efficient deeplearning model training.


Vector Search Databases

  • Milvus: Knowledgeable in building similarity search applications using this highly scalable platform.
  • Pinecone: Experience in implementing vector search for machine learning models in production.


Databases & Data Warehousing

  • Athena: Proficient in serverless queries with SQL on large-scale data directly in S3.
  • Snowflake: Skilled in utilizing this cloud data platform for scalable analytics.
  • Redis: Experienced in using this in-memory database for caching and real-time analytics.
  • PostgreSQL: Strong understanding of relational database management and development


Cloud Computing Platforms

  • AWS: Extensive experience with Sagemaker for ML model development and deployment; Redshift for data warehousing; Lambda for serverless computing; Personalize recipes for Recommender models.
  • Google Cloud Platform(GCP): Proficient with BigQuery for data warehousing; developed a Recommender system on this cloud provider.


Knowledge Graphs

  • Neo4j: In-depth experience with this graph database for building knowledge graphs and complex queries.


Data Engineering and Stream Processing

  • Apache Kafka
    • Advanced proficiency in stream processing systems for building faulttolerant, scalable real-time data pipelines.
  • Analytics Databases
    • Apache Druid: Experienced in real-time analytics with Druid, enabling interactive queries and insights on large-scale datasets.
  • Data Visualization Tools
    • ?Grafana: Experienced in deploying Grafana for comprehensive monitoring and visualization of metrics and logs across various data sources.

Work Experience

09/2023 - today
Senior Data Scientist (Freelancer)

08/2021 - 08/2023
Senior Data Scientist 
Bayerischer Rundfunk/.pub, Munich (Germany)
  • research on state-of-the-art developments in Recommender Systems in media (audio-, video- and textual content)
  • implementation of a Recommender System powering ARD Audiothek (one one Germany?s most popular audio-on-demand platforms). The deployed production model has 15% higher precision than the previous
  • NLP projects: entity recognition and redundancy removal

06/2020 - 07/2021
Data Scientist 
Yewno/Entropy387, Sarajevo (BiH)/San Francisco (US)
  • research on and implementation of graph-based stock-market prediction models
  • pricing engine implementation (achieved a 10% improvement of MAE with respect to the previous model)

06/2018 - 05/2019
Post-Doc 
Roma Tre University, Rome, Italy
  • research on Graph Morphing Algorithms
  • implementation of Graph Drawing Algorithms

04/2014 - 05/2016
Software Engineer 
Visteon (former Johnson Controls), Karlsruhe, Germany
  • (re)implementation of a testing software
  • implementation of Finite State Machines
  • code generation

Programmiersprachen

Python
C/C++
Java
MATLAB
Shell
Batch

Datenbanken

MS SQL Server
T-SQL

Einsatzorte

Einsatzorte

Deutschland, Schweiz, Österreich
möglich

Projekte

Projekte

8 Monate
2024-11 - 2025-06

Agentic AI for a DeepResearch project

To create a multi-agentic system, supported by a Knowledge Graph, that automates the process of drafting a research paper. The system used multiple experts (OpenAI models) that ?collaborated? during the process of document drafting. The whole process was supported by a Knowledge Graph out of which we extracted useful information. Technologies used for this framework: LangChain, LangGraph, Smolagents, LlamaIndex, dspy. The project involved the use of Terraform and GitHub Actions (CI/CD Pipeline), via AWS. The initial application was deployed as a Streamlit app.
6 Monate
2024-05 - 2024-10

Open source project - PulseSpotter

PulseSpotter is designed to assist journalists in the task of identifying newsworthy topics that are likely to become popular. By gathering information from various news sources and analyzing patterns over time, the system suggests emerging trends by using AI and machine learning, saving journalists time on research and deciding which stories to focus on.
MediaLab Bayern
8 Monate
2024-02 - 2024-09

RAG (LLM)

Objective: an LLM chatbot to help with HR-related inquiries.
  • Approach: I led the development of an advanced Retrieval-Augmented Generation (RAG) system aimed at improving HR data retrieval processes. This system utilizes the Milvus Vector Similarity Search Database and OpenAI?s API to efficiently source and integrate extensive HR-related data, enabling it to respond to a broad spectrum of HR inquiries. We have also attempted to fine-tune an LLM for this task and the technologies used were PEFT (LoRA), and sophisticated Prompt Engineering.
1 Jahr 1 Monat
2023-09 - 2024-09

Spectral Image Object Detection

Objective: given a spectral image that depicts the value of the sensors?s readings, classify the signals and identify novelties (anomalies).
  • Approach: Given that there was not sufficient labeled data, I had to rely on selfsupervised machine learning paradigms. To satisfy the customer?s request for fast processing, I utilized one of the YOLO architectures. The system was developed with the mmyolo framework.
5 Monate
2023-08 - 2023-12

User Grouping in a Social Network

  • Objective: To foster communication among like-minded users through dynamic group formation based on their responses to specific questions.
  • Approach: Employed vectorization and K-Nearest Neighbors (KNN) for initial grouping, with a recommendation system for refined group alignment.
5 Monate
2021-04 - 2021-08

Fraud Detection for a Buy-Now Pay-later Platform

  • Objective: Fraud detection.
  • Approach: I designed and developed a fraud detection model for a Middle-Eastern Buy-Now Pay-Later platform. A critical step for this client involved data preprocessing, during which I employed a graph-based approach to identify cliques of fraudsters. This was the first successful machine learning project for this client.
8 Monate
2020-09 - 2021-04

Next-Best Finance-Related Action

  • Objective: To advise individuals on financial actions that enhance their chances of achieving specific goals.
  • Context: For an Asian bank aiming to offer actionable advice for long-term financial planning, such as securing a home purchase in 20 years.
  • Approach: Predictive machine learning models were developed to suggest optimal actions (example: obtaining a salary increase of at least 10%).
4 Monate
2020-05 - 2020-08

Recommender for Online Betting

  • Objective: To surpass the performance of Amazon Personalize?s Hierarchical Recurrent Neural Network-based recommendation system.
  • Approach: Adopted a recommender model based on Factorization Machines, significantly improving recommendations through targeted feature engineering.
8 Monate
2019-09 - 2020-04

Debt Collection

  • Objective: To predict the likelihood of loan repayment by bank customers, aiding in their segmentation for tailored communication strategies (email, SMS, or phone calls)
  • Achievement: Created a behavioral scoring machine learning model, now incorporated into Receeve?s collection approach.
8 Monate
2019-01 - 2019-08

Sensor Anomaly Detection

  • Objective: To detect unusual behavior in a sensor-monitored network.
  • Method: Established a baseline machine learning model predicting expected sensor readings, using deviations from this model to flag anomalies. Enhanced detection accuracy with a Graph Neural Network, analyzing the connectivity between sensors and observable areas.
9 Monate
2018-04 - 2018-12

Pricing Engine for a US Dispatcher

  • Objective: To predict vehicle transportation prices, providing dispatchers with a reliable basis for pricing.
  • Achievements: Enhanced the machine learning model?s accuracy by approximately 40% through advanced feature engineering and implementing a two-tiered regression strategy, combining residual and standard regression techniques powered by XGBoost.
3 Monate
2018-01 - 2018-03

Investment Prediction

  • Objective: To forecast investors? forthcoming decisions on company investments.
  • Approach: Transformed the challenge into a graph-based problem, treating investors and companies as nodes. Developed a Graph Neural Network to predict potential investor-company connections, integrating this model into StockFink?s prediction suite.

Aus- und Weiterbildung

Aus- und Weiterbildung

3 Jahre 2 Monate
2009-11 - 2012-12

Study - Computer Science

Ph.D., The University of Konstanz (Germany)
Ph.D.
The University of Konstanz (Germany)

  • Scholarship by the German Research Foundation (DFG)
  • Work Group: Algorithmics
  • Thesis: on request

3 Jahre 9 Monate
2005-10 - 2009-06

Study - Computer Science

Bachelor of Science, The University Sarajevo School of Science and Technology, Sarajevo, BiH, The University of Buckingham, Buckingham, UK
Bachelor of Science
The University Sarajevo School of Science and Technology, Sarajevo, BiH, The University of Buckingham, Buckingham, UK

  • Major in Computer Science, minor in Electrical and Electronic Engineering

Kompetenzen

Kompetenzen

Top-Skills

Data Science Recommender Systems Algorithms

Aufgabenbereiche

  • LLM
  • RAG (Graph-RAG)
  • Agentic AI
  • Recommender Systems
  • Deep Learning
  • Graph Neural Networks
  • Pricing Engines
  • Machine Learning
  • Dimension Reduction
  • Graph Embedding and Information Visualization
  • Algorithm Engineering (large data processing)
  • Numerical Optimization

Produkte / Standards / Erfahrungen / Methoden

Profile
The candidate received his doctorate in computer science from the University of Konstanz at the age of 25. During his academic career, he worked in the areas of dimensionality reduction and graph embedding, and his work has been recognized by the scientific community. As a (senior) data scientist, the candidate focuses on LLMs/RAGs, recommender systems, knowledge graphs, and classical machine learning. His most notable work concerns the development and implementation of a recommender system for the ARD audio library. He is also a fiction author.

Computer Skills
Python
  • Numpy
  • Pandas
  • scikit-learn
  • TensorFlow
  • Keras
  • PyTorch
  • PyG
  • Networkx
  • Matplotlib/ Seaborn
  • XGBoost/ LightGBM/ CatBoost
  • SpaCy/ NLTK

Other
  • AWS: 
    • deployment
    • Lambda
    • EC2
    • ECS
  • Azure: 
    • Databricks
  • Agile Development: 
    • Scrum
    • Kanban
  • Kubernetes
  • Streamlit
  • Spark

Scripting
  • Shell
  • Batch

Typesetting
  • LaTeX

Recommender Systems
  • collaborative filtering (linear models, matrix-factorization)
  • ensemble models
  • bandit-based models
  • graph-based recommenders
  • neural-network recommenders


Software

Falcone ? a graph profiling software based on multidimensional scaling techniques, written in Java


Deep Learning Frameworks

  • TensorFlow & Keras: Extensive experience in building and deploying neural network models.
  • PyTorch: Proficient in model development and experimentation, with a focus on dynamic computation graphs.
  • PyTorch Geometric (PyG): Skilled in implementing graph neural networks for complex relational data.
  • PyTorch Lightning: Familiar with this framework for scalable and efficient deeplearning model training.


Vector Search Databases

  • Milvus: Knowledgeable in building similarity search applications using this highly scalable platform.
  • Pinecone: Experience in implementing vector search for machine learning models in production.


Databases & Data Warehousing

  • Athena: Proficient in serverless queries with SQL on large-scale data directly in S3.
  • Snowflake: Skilled in utilizing this cloud data platform for scalable analytics.
  • Redis: Experienced in using this in-memory database for caching and real-time analytics.
  • PostgreSQL: Strong understanding of relational database management and development


Cloud Computing Platforms

  • AWS: Extensive experience with Sagemaker for ML model development and deployment; Redshift for data warehousing; Lambda for serverless computing; Personalize recipes for Recommender models.
  • Google Cloud Platform(GCP): Proficient with BigQuery for data warehousing; developed a Recommender system on this cloud provider.


Knowledge Graphs

  • Neo4j: In-depth experience with this graph database for building knowledge graphs and complex queries.


Data Engineering and Stream Processing

  • Apache Kafka
    • Advanced proficiency in stream processing systems for building faulttolerant, scalable real-time data pipelines.
  • Analytics Databases
    • Apache Druid: Experienced in real-time analytics with Druid, enabling interactive queries and insights on large-scale datasets.
  • Data Visualization Tools
    • ?Grafana: Experienced in deploying Grafana for comprehensive monitoring and visualization of metrics and logs across various data sources.

Work Experience

09/2023 - today
Senior Data Scientist (Freelancer)

08/2021 - 08/2023
Senior Data Scientist 
Bayerischer Rundfunk/.pub, Munich (Germany)
  • research on state-of-the-art developments in Recommender Systems in media (audio-, video- and textual content)
  • implementation of a Recommender System powering ARD Audiothek (one one Germany?s most popular audio-on-demand platforms). The deployed production model has 15% higher precision than the previous
  • NLP projects: entity recognition and redundancy removal

06/2020 - 07/2021
Data Scientist 
Yewno/Entropy387, Sarajevo (BiH)/San Francisco (US)
  • research on and implementation of graph-based stock-market prediction models
  • pricing engine implementation (achieved a 10% improvement of MAE with respect to the previous model)

06/2018 - 05/2019
Post-Doc 
Roma Tre University, Rome, Italy
  • research on Graph Morphing Algorithms
  • implementation of Graph Drawing Algorithms

04/2014 - 05/2016
Software Engineer 
Visteon (former Johnson Controls), Karlsruhe, Germany
  • (re)implementation of a testing software
  • implementation of Finite State Machines
  • code generation

Programmiersprachen

Python
C/C++
Java
MATLAB
Shell
Batch

Datenbanken

MS SQL Server
T-SQL

Vertrauen Sie auf Randstad

Im Bereich Freelancing
Im Bereich Arbeitnehmerüberlassung / Personalvermittlung

Fragen?

Rufen Sie uns an +49 89 500316-300 oder schreiben Sie uns:

Das Freelancer-Portal

Direktester geht's nicht! Ganz einfach Freelancer finden und direkt Kontakt aufnehmen.