Sr. Machine Learning Analyst - The NPD Group (New York - USA), March 2016 - May 2017
Hadoop Big data Scientist - Intel Inc. Data Science Group (New York - USA), January 2016 - March 2016
• Developing and planning required analytic Big Data and Predictive Analytics project requirements in response to business needs in Hadoop and other relevant technologies. • Developing new predictive/analytical modeling methods and/or tools as required in Big Data with required Machine Learning algorithms with Hadoop technologies. • Working with business domain experts and application developers to identify data relevant for analysis. • In conjunction with BigData data owners and clients teams, contributing to the development and evolution for data models for analytical capabilities. • Contributing to predictive/analytical modeling architectures, modeling standards, reporting, and data analysis methodologies with Hadoop. • Contributing to recommendations on predictive/analytical modeling products, services. protocols, and standards.
Sr. Machine Learning Analyst, Data Scientist - Project Developer - Biltek Inc. (Istanbul-TURKEY), October 2011 - November 2015
1- Web-based recommendation system for travel agencies and banks with real-time processing in concurrent threads processing in Python and C++. It keeps track of customer activities who visited the webpages with banners. SVM and logistic regression were employed for predictive analytics. The recommendation system employed text processing to predict customer tendency with use of NLP and ML where ML was employed for recognizing associations and identify the possible collocated customer front-end entries. Use of Elastic-search for text patterns to be searched, aggregated & analyzed. A heuristic web crawler also developed for search in this and following project in Python.
2- Web-based predictive analytics of students’ performance in education. Due to categorical attributes to be transformed to possible associations with respective records, a newly developed clustering algorithm was employed to group them. Modeling based on optimization & classification.
3- Design of a human-computer interface (HCI), to be used in assessing medication intakes of epileptic human subjects with multi-channel EEG recordings. A new feature definition for multi-channel EEG waveforms was presented for predicting human emotions & verified for regression classifiers.
• Developing webpages & admin panels for customers and BILTEK requested by such customers as travel agencies & banks. Back-end coding with PHP & data transactions through MySql DBs on Linux Apache servers, i.e. LAMP, with HTML, CSS. PHP also used for retrieving & manipulating data from XML & RSS resources e.g. exchange rates & weather. Front-end coding with JS, jQuery, Ajax. Related JS coding to run Flash plug-ins for animations.
• Developing projects for control & communication applications & their embedded software & hardware implementations with use of C/C++ or Java techs mainly for defense & environmental monitoring with remote sensing.
Research Scientist/Engineer - Tubitak/Turksat (Ankara - TURKEY), February 2006 – October 2006
Working on payload design, navigation, control and guidance of ATAK helicopters, and communication satellites >10M USD
Varying IT firms, June 2005 - August 2010
Mentor, consultant and part-time system designer and developer for various industrial and public projects involving software and hardware development involving various DSPs (TI and NI families), Motorola, Atmel etc. microcontrollers.
Head and Assoc. Prof. in Mechatronics Engineering Department (Bursa - TURKEY), April 2013 - Oct 2015
- Teaching various graduate and undergraduate courses -
Graduate courses taught : Machine Learning and Data Mining, Big Data Management and Process Mining, Advanced Object-oriented Programming & Networking, Estimation Theory and Adaptive Filtering, Guidance, Navigation and Control, Cryptography for Secure Information Processing, Estimation and Optimal Control for Guidance & Navigation, Animative Networking and Multimedia, Neural Computation and Control, Mobile Computing Systems & Programming,
Undergraduate courses taught : Computing for Engineers, Computer Programming (C/C++), Data Structures (JAVA), Object-oriented Programming with JAVA, Introduction to Programming (C++), Introduction to Programming (JAVA),
- Principal Investigator (PI) in project ‘Guidance, navigation and control for multi-rotor unmanned air vehicles’.
Sponsor: BTU Scientific Research Programme (SRP) fund with budget 5000 Euro.
- Researcher in project ‘Design and implementation of a PLC-controlled fully automated system for inspection of reliability and consistency of a computer-aided standardized photographic imaging method in evaluation of morphometric gum reviews’.
Sponsor: Erciyes Univ. SRP Fund.
Visiting Prof./Postdoctoral Researcher - College of Engineering, Villanova University (PA, USA), February 2011 - October 2011
- Research in project ‘Navigation, tracking, guidance and control of Underactuated Autonomus Vehicles (UAV) based on sliding-mode speed controllers & Extended Kalman Filter with Bluetooth communication’.
Sponsor: USA Navy with budget > 10M USD.
Assistant Professor, Computer and Software Engineering Departments, Bahcesehir and Halic Universities, (Istanbul, TURKEY), September 2006 - September 2010
- Teaching various graduate and undergraduate courses,
- Principal Investigator (PI), in project ‘Implementation of neuro-spike-based sensory itegration of hard & soft algorithms on mini-robots’ for object recognition and localization.
Sponsor: Univ. SRP with budget 5000 Euro.
- Principal Investigator (PI), in project ‘VLSI Implementation of Multi-valued Logic Systems’ between July 2007 - July 2010.
Sponsor: TUBITAK with budget >150k Euro.
Postdoctoral Researcher - School of Informatics, Edinburgh University (UK), Department of Artificial Intelligence, February 2003 - May 2005
- Research in a European Union 5th Framework Program project ‘CIRCE
Implementation of Biomimetic Robot with Auditory Structure Inspired by Bats’. Implementation of machine learning and pattern recognition algorithms for sonar echolocation, identifying scenario targets and implementing device drivers with necessary software in C/C++ on Linux with involved memory management and parallel programming.
Sponsor: European Union EPSRC with budget > 2.1M Euro.