Warren Pfeffer, Research Fellow & Director of R&D Lab

Warren Pfeffer has been an IT professional since 1983, and has been working for financial services IT organizations since 1987.

In his current role at Oktay Technology, Warren is Director of the R&D Lab where he architects, implements, and executes proof-of-concept evaluations for Big Data platforms and Low-Latency technologies. Warren is also a research fellow at Oktay specializing in research into Big Data technologies such as Hadoop, Cassandra, Kafka, Storm, Spark, Elasticsearch and a variety of other NoSQL / MPP technologies. Warren also specializes in application performance analysis, tuning, and debugging, and in the development of benchmarking programs for evaluating system performance.

Warren has developed and maintained a number of real-time trading applications for leading investment banks on Wall Street (most recently at Credit Suisse for 13+ years and previously at Merrill Lynch for 10+ years).

At Credit Suisse, Warren made significant architectural, performance, and scalability enhancements to the core equities order management system. These enhancements included: refactoring the entire code base from 32bit Solaris to 64bit Linux; optimizing TCP and memory map I/O performance; and incorporating HA clustering. Warren was also a founding member of the Credit Suisse R&D team that successfully introduced transformational technologies such as the first Linux systems, real-time Linux, blade servers, NUMA servers, Infiniband networking, and high performance market data appliances. Warren also worked as the DevOps Manager on the Credit Suisse Trading Horizontal project were he introduced and maintained technologies to improve developer productivity such as Intel Vtune, Systemtap, Maven/Nexus, Jira, and Team City. Warren also developed a Java based Hadoop application for OMS log file mining.

At Merrill Lynch, Warren was a senior developer on the team that created the first two generations of NASDAQ trading platforms for Merrill. In addition to creating high performance messaging protocols for these C++ applications, Warren also created sophisticated parsing frameworks to transform the full NASDAQ messaging specification into Yacc grammar for code generation and to enable Lex/Yacc to be used as a dynamic parser at runtime for forms validations. Warren also developed an X based color heat-map system for monitoring the health of trading floor applications.

Warren has a B.A. in Computer Science from Thomas Edison State College and an A.S. in Data Processing from Norwalk State Technical College.