I specialize in developing innovative ML solutions for regulatory compliance challenges (see corresponding publication list). My key attributes include: - Science lead for a group of scientists focused on creating high-performing ML models for production-ready compliance applications. - Develop cutting-edge ML solutions that deliver tangible business impact. - Innovate and produce research papers of premium-tier quality.

Building computer vision / machine learning solutions to address compliance issues for Amazon products.

Contractor position (through Adecco Romania) which involves working on-site at the Google Bucharest office outsourcing a Google Research project.My main responsibilities include: - Developing machine learning solutions for automatic data classification and pattern recognition on par with state-of-the-art performance. This has resulted in 2 papers at AAAI and CVPR (see list below). - Creating automatic tools for data processing, visualization, packing and quality control. - Operating 3d surface scanning equipment.

Perform research on topics such as estimating, modeling or synthesizing the 3d human body shapes under various constraints (i.e. geometrical, scene). My top responsibilities included the following: - Design and implement methods that use techniques such as deep neural networks, GANs, variational auto-encoders, graphical models, numerical optimization, 3D geometrical modelling, kernel methods/semi-supervised learning. - Write scientific papers to be published at conferences such as CVPR, ICCV, ECCV, NeurIPS.

Project based collaboration which involved building a 3D avatar model world (football supporters) with human – computer interaction based on Microsoft Kinect and smartphone(Android & iOS). I worked in a team of 2 and we were responsible for:- developing a C# library for gesture recognition which involves Microsoft Kinect SDK- the detection of the hand which held the phone- the mapping of the 3D objects on the avatar model according to the data received from you skeleton.

The position implied teaching object orientated programming from a practical perspective (lab sessions) for first year students at Faculty of Mathematics and Computer Science. I illustrated and taught object oriented programming paradigm concepts by using the C++ language.

The work involved building a C++ library for hand gestures recognition which were performed by using a Microsoft Kinect device. Support Vector Machine model was used as a machine learner model for the project work and the following libraries were also used: openGL, openCV, openNI. For demo, click the link, https://sites.google.com/site/gesturerecognitionbii/

Performed research work over Magnetic Flux Leakage data towards building a machine learning model for detection/mapping of pipe defects.