Jeffrey D. Oldham

144 Wyandotte Dr, San Jose, CA 95123-3727


Software Development and Algorithmic Research

Experience and Skills

Software Engineer:
CodeSourcery, LLC (June 2000-). Designed, specified, and implementing the C++ VSIPL++ toolkit supporting vector, signal, and image processing in both serial and distributed environments. Completed USAF OSD Phase 1 SBIR to implement VSIPL++ extensions to support multiple processors and SIMD instructions. Wrote Advanced Linux Programming and POOMA user manual. Specified and developed QMTest test regression tool. Wrote run-time C++ library. Optimized financial programs. Performed patent law work. Compiler development including writing C++ ABI test library, optimization work, bug fixes.

Assistant Professor:
Trinity University Computer Science Department (August 1999-May 2000). Taught courses in introductory programming (two semesters), operating systems, and simulation. With 44 students, designed and implemented sales tax server prototype. With 12 students, developed models and simulated investment accumulation and contraction.

Visiting Assistant Professor:
Dartmouth College Computer Science Department (Spring 1999). Taught advanced algorithms.

AT&T Labs Researcher:
AT&T Research Labs (Summer 1998). The proposed HTTP 1.1 protocol permits persistent connections between servers and clients. Designed and evaluated policies to predict when to terminate persistent server-client connections.

Stanford University Ph.D. Candidate:
Computer Science Department (September 1992-August 1999). Designed theoretically-fast algorithms for multicommodity and generalized flow problems; multicommodity flow implementation was several orders of magnitude faster than previous implementations. Designed and experimentally evaluated faster algorithms to value Asian options. Implemented DNA sequence alignment algorithm, used in MIS 214 course. Instructor for compiler course and programming languages courses. Teaching assistant for intensive C++ and compiler optimization courses.

Recent Software, Specifications, and Publications

Specified and implementing the C++ Vector, Signal, and Image Processing Library for use in both serial and distributed environments using intuitive, algorithmic syntax. Wrote 250+ page specification incorporating the best current software research. Work funded by the Department of Defense High Performance Embedded Computing Software Initiative coordinated through MIT Lincoln Laboratory (

Multicommodity Flow:
Combinatorial approximation program was faster than its closest competitor by several orders of magnitude.

Advanced Linux Programming:
Book describing UNIX and Linux-specific library programming interfaces. Joint work with Mark Mitchell and Alex Samuel.

Developed and maintained open-source C++ software toolkit for parallel scientific computation. Users include nuclear weapon physicists at Los Alamos National Laboratory.

POOMA User's Manual:
User manual for the open-source POOMA C++ toolkit for parallel scientific computation.


Ph.D. in Computer Science:
Stanford University, August 1999. Thesis title: ``Multicommodity and Generalized Flows: Theory and Practice.'' Development and implementation of algorithms to solve multicommodity and generalized flow problems. In the first half, I converted theoretically fast but, in practice, slow algorithms into an algorithm that was several orders of magnitude faster than its closest competition. In the second half, I developed the first combinatorial algorithms to solve many generalized flow problems.
M.S. in Computer Science:
Stanford University, June 1996.
B.A. in Computer Science:
Rice University, May 1991. Summa cum laude. National Merit Scholar. Tau Beta Pi.