Naoki Tanaka

• Served as Oracle Linux's primary SELinux developer and Mandatory Access Control (MAC) subject-matter expert, advising engineers and technical leadership across multiple organizations on policy design, confinement strategy, and security troubleshooting.
• Developed and maintained SELinux policies for the default targeted policy as well as minimum and highly specialized Multi-Level Security (MLS) variants.
• Identified and resolved SELinux defects across Oracle Linux, including uncovering flaws in internally maintained test suites, improving both product quality and validation coverage.
• Initiated upstream contributions to Fedora's selinux-policy project, with 18 pull requests merged upstream, becoming the project's #19 all-time contributor — notable as the project is primarily maintained by Red Hat and Fedora engineers.
• Contributed fixes and enhancements to firewalld and OpenSSL components within Oracle Linux, addressing security, reliability, and integration issues beyond the SELinux subsystem.

• Developed Verified Boot for Kernel Zones and LDoms (Oracle VM Server for SPARC) on Solaris in C to detect accidental or malicious modification of kernel modules before loading by verifying factory-signed signatures.
• Designed and developed audit of Solaris Verified Boot events in C to generate audit trails when inappropriately-signed kernel modules are to be loaded.
• Integrated key components for multi-factor authentication such as OpenCA Research Labs' OCSP Responder and LibPKI into Solaris.
• Developed a prototype TPM 2.0 driver for Solaris in C to eliminate dependencies on obsolete cryptography algorithms such as SHA-1 used in TPM 1.2 standard.

• Designed a rational approach based on actuarial methods to encourage appropriate information sharing inside a virtual organization.
• Developed a discrete event simulator for the proposed approach in C++ with Boost C++ Libraries and its GUI interface for Mac in Objective-C with Core Plot framework.
• Designed and implemented with Node.js and MongoDB an online business simulation game to evaluate my approach.
• Conducted experiments using Amazon Mechanical Turk, and confirmed that my approach could benefit an organization even when human decisions are involved.
• Designed game theoretic and decision theoretic risk token allocation mechanisms for risk-aware authorization.

• Added new features used by machine learning algorithms for feed ranking using PHP, C++, and Hive; one ranked #1 in importance for feed ads ranking.
• Developed a web interface using PHP and JavaScript that makes it possible to holistically analyze ads data taken from multiple data sources.

• Interned under the A*STAR Graduate Academy's A*STAR Research Attachment Programme in Singapore for one year.
• Proposed and analyzed using C++ with Boost C++ Libraries a decision framework that enables a virtual organization to select an optimal portfolio of risky data accesses that will maximize the benefit subject to a given risk budget.

• Interned as one of the inaugural fellows of the NYC Turing Fellows Program, which seeks to match top computer science and engineering students with outstanding summer internships at leading NYC startups.
• Proposed effective statistical and visualization methods using Python, PostgreSQL, and Gnuplot for analyzing TV ad campaign structures from huge amount of set-top box data.

• Developed a distributed non-intrusive load monitoring program in Java using genetic algorithm with JGAP framework and dynamic programming.

• Led discussion sections of an introductory Computer Science & Java class for CS majors.

• Managed various company-wide IT projects.

• Located several active compounds by applying data mining techniques such as random forests written in C++ to huge volume of assay data stored in Oracle DB using HPC cluster and Grid computing environments such as Platform LSF and Oracle Grid Engine.
• Constructed an LDAP+Kerberos centralized single-sign-on authentication system by Redhat Enterprise Linux in the heterogeneous environment of Linux and Windows.

Thesis: Sustainable Approaches to Ad-Hoc Information Sharing for Virtual Organizations [PDF]

Advisor: Professor Marianne Winslett

GPA: 4.0 (Cumulative)

Thesis: A Method for Analyzing Metabolic Networks Based on Gene Essentiality in a Focus of Compounds

GPA: 4.0 (Cumulative)

Thesis: A Systematic Method for the Experimental Data Analysis of the Gene Disruption Strains of E. coli

GPA: 3.87 (Cumulative), 3.94 (Major)

• Kusumoto Award (Top Student)