Printable version (pdf) Click here to go back to
homepage
JAYADEV
JAYARAMAN
2126
Oak Street, Los Angeles, CA 90007 Email: jdevjram@gmail.com
Phone:
(310) 729-7230 Website: www-scf.usc.edu/~jjayaram
OBJECTIVE: A challenging position as a
Software/Research Engineer
Programming Background
& Software Skillset
· Programming languages: C/C++,
JAVA, Python, PERL, Javascript, PL/SQL, Verilog HDL
· Packages: MATLAB, MySQL, Lex & Yacc, Altera Quartus II
· Natural language toolkits: GIZA++, Carmel, Tiburon, SRILM
· Operating Systems: Microsoft Windows,
GNU/Linux (Debian/Ubuntu)
Project Experience
1.
“Linear Time
Decoding in Machine Translation” – Directed Research (Jan ‘10 –
present)
Advisors:
Dr. D. Chiang & Dr. L. Huang, ISI Natural Language
Group, Marina Del Rey, CA
Project Accomplishments:
· Adapting
SCFG-based left-to-right target generation algorithm (T. Watanabe, 2006) to
work with generic SCFGs.
· Improving
decoder performance using cube-pruning (Chiang, 2007); investigating
left-corner parsing variants.
2.
“A Word-based
French-English Machine Translation (MT) System”
(Oct ‘09 – Dec ‘09)
Advisors: Dr. D. Chiang & Dr. L.
Huang, ISI Natural Language Group, Marina Del Rey, CA
Project Accomplishments:
· Built and
evaluated a MT system for French to English translation based on IBM
word-alignment models.
· A stack decoder
incorporating beam-search was developed and used to generate candidate
translations.
3.
“Causal link
detection in data using Monte Carlo methods”
(Mar ‘10 – May ‘10)
Advisor: Dr. F.
Sha, Dept. of CS, Univ. of S.
California, Los Angeles, CA
Project Accomplishments:
· Implemented a
method for weighting hypotheses of Bayesian Network structure, given
sparse-data.
· Monte-Carlo
sampling of BN orderings was used to weight edge and path features given a
dataset.
4.
"Feature-phrase
directed extraction of sentiment from text”
(Jan
‘09 – May ‘09)
Advisor: Dr. V.
Gopalakrishnan, Dept. of CSE., NIT Trichy, India
Project Accomplishments:
· Designed and developed methods for unsupervised
learning for Hidden-Markov-Model based PoS tagging.
· A PoS tagger was built and used to build a MLE
(of word mixture-models) based feature-phrase extractor.
Education
M.S.
Computer Science (Specialization: Human
Language Technology) Aug
‘09 – Dec ‘10
University of
Southern California, Los Angeles, CA GPA: 3.74/4.00
· Directed Research experience in NLP;
Secured A grades in advanced level NLP courses
B.Tech. Computer Science &
Engineering Jul ‘05 – May ‘09
National Institute of Technology (formerly Regional Engineering College), Trichy, India GPA: 7.71/10.00
Achievements
· School-first: 98% in Computer Science – Class XII final All-India Senior School Certificate Exam
(2005)
· General
proficiency: 95.6% overall – Class XII final All-India Senior
School Certificate Exam (2005)
· Obtained state rank 95 & admission to NIT,
Trichy – All India Engineering Entrance Exam (AIEEE
– May 2005)
· 1056
points
in
3 online Single Round Programming Contests, on Topcoder
(handle: thejdev; www.topcoder.com).
Volunteer Activities
· Coordinator, Org. Committee: ‘Vortex 2008’,
a national-level annual Technical Symposium, CSE Dept., NIT, Trichy, India
· Coordinator, Printing Committee: ‘Festember 2006’, a national-level
inter-collegiate cultural festival, NIT, Trichy, India
REFERENCES: Several excellent references
furnished upon request. (Also, please see my website for further
details.)