Machine Learning Engineer/Data Scientist2017 - PRESENTToptal and Toptal Clients
Technologies: Python, TensorFlow, Scikit-Learn, OpenCV
- Interviewed Toptal candidates as a technical screener for artificial intelligence and data science specializations.
- Developed a convolutional neural network for sound classification.
- Developed a computer vision neural network model for laboratory analysis.
- Used Spark to model genetic data.
- Developed predictive models for customer churn.
Software Development Engineer2017 - 2018Amazon
- Developed services for global fashion retail.
Research Scholar2015 - 2016Drexel University
Technologies: Java, Python, Machine Learning
- Brought in as a visiting scholar in the Department of Computer Science, Database Group.
- Researched mining and modeling rank and preference data.
- Implemented a Java library for handling and mining rank and preference data.
Research Engineer2012 - 2015INRIA
Technologies: Java, C/C++, Python, MATLAB
- Parallelized machine learning algorithms (SVM, affinity propagation, neural networks, etc.).
- Conducted high-performance computing. Adapted algorithms for a large-memory (8 TB RAM) NUMA architecture, on a low level (in C/C++), with awareness of cache processes, memory blocks latencies, and exact control over assignment of threads to cores.
- Worked on a semantic web project (RDF, Wikidata mining). Implemented a crawler and category recommender system for Wikipedia.
- Managed big data and ETL using SAS.
Assistant Professor2011 - 2012Metropolitan University
Technologies: Java, Web, UML
- Taught at undergraduate and graduate level, including courses in web systems and applications, distributed systems, and information system design.
- Developed the university information system and business process management workflow.
Post-doc Researcher2010 - 2011University of Eastern Piedmont
Technologies: Java, C/C++, MATLAB
- Contributed to a European Space Agency project on using computer vision for Mars lander navigation.
- Implemented a real-time system in C/C++ with OpenCV. Analyzed the image stream from the camera attached to Mars Lander bottom in order to calculate the position of the lander (coordinates, altitude, attitudes).
- Tracked features in the video. Filtered the output and combined information with other sensors (lidar, inertial measurement unit).
- Modeled the Mars surface in Java 3D.
- Built a landing simulation in C/C++ and MATLAB.
Ph.D. Student2006 - 2010University of Turin, Department of Computer Science
Technologies: Java, Weka
- Did my doctoral studies in the field of data mining and machine learning.
- Worked with text mining on local government data and document classification.
- Researched sequential pattern mining; recognizing users by keyboard strokes.
- Studied log mining, modeling and predicting usage of a grid computing network.
- Worked with graph analysis and complex systems analysis, as well as visualization of complex data.
Software Engineer2001 - 2006RCUB
Technologies: Java, Web, UML
- Joined RCUB (Belgrade University Computer Center), where the best software engineering students, after graduation, are offered an opportunity to work on commercial and research projects in an inspiring environment.
- Designed and architected Version 2 of a network (WAN) monitoring system used in several huge organizations and companies (banks and telecommunications).
- Designed, implemented, and led the team of the hospital information system currently in use in about half of the hospitals in Serbia.
- Gained experience in all aspects of implementing large information systems: from interviews, specifications, and UML model, to database design, back-end business logic, to front-end, web (including web design), and stand-alone clients (Delphi at Java Swing at that time).
- Also worked on a government information system, fleet management system, and more.