Kirill Trapeznikov +1-440-***-****
ab9gl3@r.postjobfree.com
http://blogs.bu.edu/ktrap/aboutme/
*** ****** **., *********, ** 02445
citizenship: U.S. citizen
Objective
A research engineering position in an exciting multi-disciplinary environment.
Areas of Specialization
Machine Learning: supervised, semi-supervised and unsupervised algorithms, generative and discriminative
•
methods, robust classification, cost sensitive learning, feature extraction and dimensionality reduction
• Statistical Signal Processing: recursive estimation, image processing and reconstruction, inverse problems,
detection theory, basic computer vision
• Optimization Methods: convex, non-convex, online, bayesian
Software Skills: MATLAB (MEX interface), Python, C/C++, LTEX, Cadence (IC,PCB), Verilog
A
Other Skills: Knowledge of analog and digital circuit design. Use of engineering lab equipment: spectrum
analyzer, oscilloscope, function generator, various soldering tools, basic optical equipment. Personal computer
support and repair.
Education
Boston University, Boston, MA
Expected, Spring/Summer 2013
Doctor of Philosophy Candidate, Electrical Engineering
Thesis Title: ”Machine Learning on a Budget”
December 2010
Master of Science, Electrical Engineering. GPA: 3.95/4.00
September 2007
Bachelor of Science, Electrical Engineering. GPA: 3.86/4.00
Research and Professional Experience
Dept. of Electrical and Computer Engineering, Boston University, Boston, MA
September 2008 - Present
Graduate Research Assistant, Information Sciences and Systems Lab
Research in machine learning and statistical signal processing, theory and methods:
• Active learning, boosting methods, multi-stage sequential decision systems, cost-sensitive and budget
constrained classification
• Applications to explosive detection systems as part of DHS research center on Awareness and Localization
of Explosive Related Threats.
Research Advisors: Venkatesh Saligrama, David Castanon.
Sandia National Laboratories, Solar Technologies, Albuquerque, NM
Summers: 2008, 2009; Part-time: 2010 - 2012
Graduate Technical Intern
Work on concentrated solar power dish systems:
• Automated mirror facet alignment and surface characterization using fringe reflection techniques.
Development and implementation of algorithms and GUI in MATLAB and C.
• Circuit design and PCB layout for a heat engine simulator system
Biomimetic Systems, Cambridge, MA
Summer 2006
Technical Intern
Validation and testing of hardware and algorithms for an acoustic direction finder system (gunshot localization).
Selected Publications
K. Trapeznikov, V. Saligrama, D. Castanon. ”Multi-Stage Classifier Design”, Machine Learning, November
•
2013
K. Trapeznikov, V. Saligrama, ”Supervised Sequential Classification Under Budget Constraints”, Int. Conf. on
•
Artificial Intell. and Stats., April 2013, (oral, 10% acceptance rate)
1
K. Trapeznikov, V. Saligrama, D. Castanon. ”Multi-Stage Classifier Design”, Asian Conf. on Machine
•
Learning, November 2012, (oral)
K. Trapeznikov, V. Saligrama, D. Castanon. Two Stage Decision System, IEEE Stochastic Signal Processing
•
Workshop, August 2012
K. Trapeznikov, V. Saligrama, D. Castanon. ”ActBoost: Active Boosted Learning”, Int. Conf. on Artificial
•
Intell. and Stats., April 2011.
C.E. Andraka, J. Yellowhair, K. Trapeznikov, J. Carlson., B. Myer, K. Hunt. ”AIMFAST: An Alignment Tool
•
Based On Fringe Reflection Methods Applied To Dish Concentrators”, J. Solar. Energy Eng 2011.
C.E. Andraka, S. Sadlon, B. Myer, K. Trapeznikov, C. Liebner. “Rapid Reflective Facet Characterization
•
Using Fringe Reflection Techniques”, ASME Energy Sustainability 2009, July 2009.
Invited Talks
Supervised Sequential Classification Under Budget Constraints, Graduation Day Talk, Information Theory
•
and Applications Workshop, San Diego, 2013
Multi-Stage Decision System, 8th Algorithm Development for Security Applications Workshop, Boston, 2012
•
Workshop Organization
Co-organizer: Int. Conf. on Machine Learning 2013 Workshop on Machine Learning with Test-time budgets
•
Poster Presentations
Sequential Decision System Design, Workshop on Multi-Trade Offs in Machine Learning, Conference on
•
Neural Information Processing Systems, Lake Tahoe, Nevada, 2012
Multi-Stage Classifier Design, Research and Industrial Collaboration Conference (RICC), at Awareness and
•
Localization of Explosive Related Threats (ALERT) DHS Center of Excellence, October, Boston, 2011
Active Boosted Learning, Boston University Science Day, 2011
•
Active Boosted Learning, Research and Industrial Collaboration Conference (RICC), at Awareness and
•
Localization of Explosive Related Threats (ALERT) DHS Center of Excellence, October, Boston, 2010
Related Coursework
Statistical Pattern Recognition, Optimal Filtering and Recursive Estimation, Linear and Non-Linear Optimization,
Image Reconstruction and Restoration, Digital Signal Processing, Information Theory, Stochastic Signals and
Systems, Wireless Communications, Analog and Digital VLSI Circuit Design, Introduction to Photonics
2