Seyhan Karakulak

Jump to other IT Society Websites:
Seyhan Karakulak portrait

Seyhan Karakulak

Technologist

Affiliation: Insight Data Science

Contact Information

8520 Costa Verde Blvd 3306,

San Diego, CA 92122

Research Interests

  • Pattern Recognition
  • Coding Theory
  • Statistical Learning and Inference
  • Detection and Estimation
  • Source Coding
  • Communications
  • Complexity and Cryptography
  • Quantum Information Theory
  • Sequences
  • Coding Techniques
  • Communication Networks
  • Shannon Theory

Biography:

 

Seyhan Karakulak, PhD

EXPERIENCE


15 years of cumulative industry and research experience in the areas of system architecture for solid- state storage devices, information theory, communication theory, error correction coding, signal processing, machine learning, and statistical learning for magnetic recording and solid-state storage systems. Have been involved with several projects as a system architect that resulted in successful ASIC implementations.

Fellow, Insight Data Science, New York City                                                                                          Jan 2020 - Present

  • Consulted for Lolli, a shopping rewards application company and helped them develop models to detect fraudulent behavior. Built supervised learning methods using logistic regression and neural networks to form a user reputation score based on the imbalanced data set with few instances of labeled fraud data.

Treasurer and the World Traveler, IEEE Data Storage Technical Committee                                   Jan 2019 - Present

  • Took a sabbatical year from January 2019 to December 2019 and volunteered for IEEE DSTC. 

Technologist, Western Digital, San Diego                                                                                         June 2016 - Dec 2018

  • Designed novel algorithms for next generation flash based solid state data storage systems to qualify performance and reliability using soft decoding and led the qualification team to incorporate these enhanced methods to the existing methods to improve system performance in one to five orders of magnitude in signal-to-noise ratios and to extend the lifetime of drives. 

  • Created soft decoding spec based on certain optimization criteria for low density parity check codes that resulted in successful ASIC implementations.

  • Co-Inventor of highly cited 9 granted patents in the data storage area.

Technologist/Principal Engineer, HGST Incorporated, San Diego                                               Sept 2013 - June 2016

  • Introduced clustering algorithms based on min-average and min-max criteria for read level profiling to improve the bit error rate in a NAND based SSD flash. This work has been published in Globecom 2016. 

  • Characterized several generations of flash channels and implemented novel detection, coding and decoding algorithms based on maximum a-posteriori detection, Hidden Markov Models, Viterbi Algorithm, soft output Viterbi Algorithm (SOVA), Expectation-Maximization, and belief propagation for soft decoding.

Systems Architect, STEC Incorporated, San Diego                                                                          April 2010 - Sep 2013

  • Designed qualification methods for decoding algorithms in flash and developed inter-cell interference algorithms for soft decoding of LDPC codes based on SOVA and BCJR type algorithms.

Intern, Qualcomm Incorporated, San Diego                                                                                     June 2007 - Sep 2007

  • Implemented and analyzed link and system simulations for Ultra Mobile Broadband project.

 EDUCATION


PhD,  MS in Electrical Engineering (Communication Theory and System), University of California, San Diego, 2007, 2010

  • Dissertation Title: “From Channel Modeling to Signal Processing for Bit Patterned Media Recording ”

    • Dissertation research involved channel modeling, detection, equalization, and information theoretic limits for bit-patterned media recording channels.

    • PhD co-advisors: Prof. Jack Keil Wolf and Prof. Paul H. Siegel

    • Relevant coursework: Information Theory, Digital Communication Systems, Algebraic Coding, Probabilistic Coding, Source Coding, Digital Communication, Statistical Learning, Probability and Statistics, Random Processes, Image Processing

BS in Electronics and Communications Engineering, and Mathematics (Double Major), Istanbul Technical University 2003, 2004