Developed an NMR Signal homogenization algorithm using Style Transfer techniques, for inhouse Oil and Gas Industry sensors.Developed an Inline NMR Spectrometer for non-stop measurement of nutritional facts in a food factory. R&D project requested by one of the world's leaders snacks manufacturers, deployed in Houston TX.

Tech Coverage: Deep Learning, Computer Vision, Activity Recognition in Thermal Imaging, Healthcare care-plans recommendation systems, CNNs, LSTMs, Classic Machine Learning Algorithms (Random Forests, etc.).· Design development deployment of a web-based engine for healthcare service recommendations on elders. This engine is first-of-its-class, and being deployed in two major Japanese cities in association with one of the largest Japanese health services provider. · Design, development, deployment of an end-to-end solution for elders monitoring at home and nursing-care facilities using deep learning algorithms in Far Infrared sensors feed. App Thermix available at the App Store, data at http://thermset.activityrecognition.com. This project won Best overall and best software at FLIR Bring the Heat San Francisco 2015.· Design, development, and deployment of an automatic video editing social network using Deep Learning. The app is free to download in the App Store under VDO.

Research work under the tutelage of Dr Francois Bremond, for team STARS. Detecting people in RGBD data for long term people tracking and events detection using Deep Learning, in the French Institute for Computer Science Research (INRIA). Contribuing to Dem@care (EU joint project for healthcare long term analysis) and Toyota Research Japan.

Final work: (March 2014) Deep Learning & Activity Recognition in sports. Based on tennis broadcasts, We applied deep learning to get scene semantics. Keywords: Machine Learning, Deep Learning, Convolutional Neural Networks. Advisors: Guido T. Pusiol, Ph.D. (Stanford University, US), Daniel E. Fridlender, Ph.D. (FaMAF - UNC, AR).

Designed, implemented and tested a self-contained anti-piracy library/protocol for Windows Applications, currently being used in Spinlock’s software. The protocol is based on AES and md5 hashing cryptography techniques. (C#) · Designed implemented and tested a self-contained protocol for auto-discovery of devices and result notification in an Enterprise network environment, using network and application layer protocols. (C#,C/C++) · Implemented and migrated the USER APPLICATION of NMR Spectrometers from Parallel/Serial interface to USB. (C#,C/C++) · Designed several Process control and Quality Control applications for company use. · Worked in the design team of the new user application of Spinlock’s NMR spectrometers (NMR4) backend with novel Pulsed Excitation console.