software engieer
Resume:
Yinglong Xia
Email: ********@***.***
http://www-scf.usc.edu/~yinglonx
Work Address
LosAngeles, CA 90089
Home Address
LosAngeles, CA 90007
Objective______________________________________________________
Doctoral candidate in Computer Science and Engineering, expecting to graduate in August 2010,
with solid background in both high performance computing and statistical machine learning, is
seeking for a challenging research and development full-time position in industry. I would like to
offer my research experience in parallel intelligent algorithm design and optimization on various
multicore/manycore and high performance computing platforms.
Education______________________________________________________
Ph.D. candidate in Computer Science (Parallel Computing) GPA: 3.8/4.0
Computer Science Department, Viterbi School of Engineering, University of Southern
California, LosAngeles, CA (Expected to graduate in August 2010)
Coursework: High Performance Computing, Parallel Computing, Computer Architectures
Dissertation: Parallelism for Probabilistic Graphical Models at Multiple Granularities
M.S. in Computer Science (Pattern Recognition) GPA: 90.3/100, 3rd out of 100
Department of Automation, Tsinghua University, Beijing, China (July 2006)
Coursework: Pattern Recognition, Statistical Machine Learning, Probabilistic Graph Model
Thesis: Statistical Machine Learning Algorithm Study on Distributed Heterogeneous Data
B.S. in Computer Engineering (Automatic Test and Control) GPA: 85.3/100, 1st out of 97
School of Automation Engineering, University of Electronic Science and Technology of
China, Chengdu, China (July 2003)
Coursework: Database, Signal and System, Electronic Measurement, Control System
Technical Skills_________________________________________________
Architectures/Platforms
Various parallel architectures, e.g., Intel® Xeon (Nehalem, Clovertown,
Dunington), AMD Opteron (Barcelona, Shanghai, Istanbul), STI Cell
BE, Sun UltraSPARC T1 and T2, nVidia GPGPU and clusters.
Languages
C/C++ with various parallel extensions, MPI, OpenMP, Shell scripts,
Cilk, Charm++, CUDA, Matlab, SQL, HTML, PHP, Java Script, CSS
Software Packages Intel® Vtune, Intel® Thread Checker, AMD CodeAnalyst, Sun Studio,
Visual Studio, Cachegrind/Valgrind, TotalView, BNT, Apache HTTP,
Emacs, LaTex, MS Office, Dreamweaver,WordPress
Operating Systems
Linux, Solaris Unix, Cluster Rocks, MSWindows
Research Experience_____________________________________________
20 publications including 5 journal papers (for details, see http://www-scf.usc.edu/~yinglonx)
University of Southern California, LosAngeles, CA August 2006 - present
Research and teaching Assistant
Explored a hierarchical scheduling scheme for the DAG structured computations on the
2
Affiliations/Activities____________________________________________
Sun UltraSPARC T2, a manycore high throughput processors (IPDPS '10)
Developed a parallel graph traversal solution called topologically adaptive BFS on
general-purpose multicore processors, such as Intel® Nehalem processors. (PDCS '09,
Best PaperAward)
Identified node level primitives with data parallelism for exact inference and developed
parallel algorithms for the kernels of exact inference. (IEEE Trans. on Computers,
SBAC-PAD ‘07)
Study a software lock-free collaborative scheduler for DAG scheduling on cache-based
multicore systems, e.g. Intel® Clovertown and AMD Barcelona.
Proposed junction tree decomposition algorithm for belief propagation on multicore
clusters. Reduced the number of communications rounds to a constant. (IPDPS '08)
Designed a centralized software scheduler for exact inference, optimized for the Cell
BE architecture, a heterogeneous multicore processor. (SC '08, JPDC)
Proposed a scalable implementation for converting Bayesian networks to junction trees
on Linux clusters. (ParCo '07)
Worked with colleagues on parallelization of belief propagation on nVidia GPGPU.
Proposed a method for processing multiple tasks in a single kernel function.
Tsinghua University, Beijing, China September 2003 - July 2006
Proposed a Markov blanket based algorithm for feature selection and ranking, which
involves probabilistic inference, collective learning, importance sampling and Bayesian
network structure learning.
Studied the fault-tolerant expectation maximization (FEM) algorithm for estimating
parameters of distributed mixture models. (DMIN '05)
Proposed the mixture random effect model (MREM) for exploring intrinsic heterogeneity of
clinical experiments from distributed medical information. (ICDM '05)
Participated in the research of the MS learning method for learning mixture models from
meta-data in medical literatures.
University of Electronic Science and Technology of China September 1999 - July 2003
Participated in the research of an Internet based embedded automated test system.
Team leader for developing an embedded information management system on Intel®
StrongARM based MSWinCE mobile platforms. (published in EST, China)
Isvision Technologies Company, Chengdu, China June 2001- August 2001
Software Engineering Internship
Participated in the development of a face-recognition system for a security product.
Affiliations/Activities____________________________________________
Student Member of the Association for Computing Machinery (ACM)
Member of Intel® Academic Community Program
Member of HPC Professionals Group
Member of Sun Developer Network
Cyber Co-chair for International Conference on High Performance Computing (HiPC '09),
Cochin, India, 2009
Web master for Southern California Smart Grid Research Symposium (SoCalSGS '09), Los
Angeles, 2009
Cyber Co-chair for International Conference on High Performance Computing (HiPC '08),
Bangalore, India, 2008
Administrator of the P-group heterogeneous cluster,USC
Student Party Branch Secretary, Tsinghua University (2004~2005)
President of School Student Union, School of Automation, UESTC (2001~2002)
Contact this candidate