Data Ninja Services is a member of the text analytics industry.  Our mission is to enable our customers to build content-intelligent applications using free texts and to enable data scientists to explore the rich semantics of structured data.  We are known as a market leader for price and performance in returning relevant content with a high degree of precision and accuracy.  Our results are aggregated by application developers and become a valuable part of their smart applications.

We are an experienced team of technology pioneers, scientists, and engineers with more than 80 years of combined experiences in both large corporate research environments and startups. We have built large-scale commercial content analytics and machine learning services for Yahoo, AOL, Visa, Lucent, Sprint, and NTT DOCOMO. As a group, we hold more than 50 patents in the areas of machine learning, data mining, content analytics, and mobile services.  We have directly contributed to several successful startup exits.

We are very flexible and work with our customers to create custom solutions. While we keep exploring and inventing, we see our customers as indispensable part of that process.

We are based in Palo Alto, California.

Data Ninja by Docomo Innovations

Meet the Team

Pero Subasic

Vice President

Before joining DOCOMO Innovations, Inc., Pero led R&D and technology teams to deliver cutting-edge technologies at Justsystem, Cadence, Yahoo, and AOL Advertising.  He pioneered Fuzzy Semantic Typing approach for sentiment analysis and holds more than 10 patents in information management related to applications of Machine Learning and Artificial Intelligence to CAD/CAE, Computational Linguistics and Information Retrieval. His teams delivered innovative, practical technology solutions leveraging large-scale distributed computing platforms with dramatic impact to key business metrics. Pero holds an Ph.D. degree in Systems and Information Engineering from Yamagata University, Japan.

Trung Diep


Trung heads the engineering team responsible for delivering the quality and performance of the Data Ninja services.  Prior to joining Docomo Innovations, Trung has previously worked at Intel, Mercury Interactive, Rambus, and Broadcom.  He received his B.S. and B.A. degrees in Electrical Engineering and Computer Science, respectively, from Rice University and M.S. and Ph.D. degrees in Computer Engineering from Carnegie Mellon University.  His research interests span from hardware, such as processor performance modeling and simulation, to software, such as cloud computing technologies, and particularly on the meta layer in which hardware interacts with software.  Trung has been granted more than 10 patents covering the areas of branch prediction, multicore arbitration and scheduling, user-level threading, cache memory, memory wear-leveling, and memory encryption.

Hongfeng Yin


Hongfeng focuses on research and development of entity extraction, knowledge engine, semantic engine technologies and services with artificial intelligence, natural language processing and machine learning approaches. Previously, Hongfeng co-founded Yebol Corporation, which developed semantic search engine technologies and services.  Prior to Yebol, Hongfeng was senior data scientist of Yahoo! data mining team for 5 years, where he developed behavioral targeting technologies and platforms with hundreds of millions in revenue.   Hongfeng has a Ph.D. degree in computer science from Concordia University, Canada.

Sayan Mukherjee


Dr. Mukherjee has worked at Bell Laboratories, Marvell Semiconductor Inc. and SpiderCloud Wireless Inc. Dr. Mukherjee has over eighty publications in journals and conferences, and has been awarded fourteen patents. He has been a Senior Member of the IEEE since 2005. Dr. Mukherjee received his M.S. and Ph.D. from Cornell University, Ithaca, NY in 1994 and 1997 respectively for work on training artificial neural network models for estimation and prediction applications. Prior to joining the Data Mining group at DOCOMO Innovations, Inc., Dr. Mukherjee was in the Wireless Networking group, where he authored the book Analytical Modeling of Heterogeneous Cellular Networks: Geometry, Coverage, and Capacity.

Hsin-Tai Wu


Hsin-Tai is an algorithm researcher, software engineer and data scientist.  He has previously taught undergraduate math and programming courses in Taiwan and US and likes to apply math to real world problems.  He has worked in depth in cryptography, GIS/GPS related algorithms/implementations, IOT, image recognition, data compression, text mining and machine learning in the past 15 years.  He co-founded two startups in the past. Prior to joining DOCOMO Innovations, Inc., Hsin-Tai worked in Foxconn and TCL developing natural language understanding related technologies.  Hsin-Tai received his Ph.D. in mathematics from UCLA.

Yas Naoi


Yas has been working for DOCOMO Innovations, Inc. and NTT for more than 20+ years in a variety of positions. His specialties include Cloud Computing, Systems Architecture Design, Agile Software Development, CI (Continuous Integration), Networking building Infrastructure, and Data Center. For DOCOMO Innovations, Inc., Yas develops and integrates Cost Visualizer for AWS cost analytics and Data Ninja Services applications on the AWS marketplace, Amazon API Gateway, and Mashape platforms. He leads DevOps for our Big Data analytics platform development and our cloud and IT infrastructure management. As an official developer, he has been granted committer privileges in the Drupal community. He holds 10 registered patents for mobile instant messenger and IoT systems. Yas Naoi graduated from Sophia University in Japan.

Maroof Khan


Presently, Maroof is working on knowledge acquisition to enhance the DOCOMO Innovations, Inc., Data Ninja services. Before joining DOCOMO Innovations Inc., he utilized statistical methods for doing predictive analytics at Intel; worked on Data modeling and generated algorithms to solve a myriad of problems at Kaiser Permanente; and built a portfolio of Machine Learning and NLP projects that showcased his technical abilities and creativity.

Maroof received his Ph.D. from Purdue University with a degree in Physics. His research focused on optical telecommunications.  His work spanned from working on the theoretical underpinnings of coupling in resonator arrays which lead to simulations and ultimately the building and testing of the first silicon based Arbitrary Waveform generator. He has over 10 publications in journals in silicon photonics.