Signal Processing Assistant

Kaushik Subramanian

**C Cedar Lane, Mobile: +1-650-***-****

Highland Park, NJ - 08904, E-mail: ********@****.*******.***

Information

www: http://www.ece.rutgers.edu/ kausubbu

USA

Rutgers, The State University of New Jersey Graduation in May 2010

Academic

Master of Science in Electrical and Computer Engineering (ECE) GPA - 3.57/4

Qualifications

Anna University, India July 2004 - May 2008

Bachelor of Engineering in ECE GPA - 72/100

Present - Project Assistant at the Real Life Reinforcement Learning Lab (RL3 ), Rutgers.

Research

Experience

Intern at RWTH Aachen University, Germany June 2009 - August 2009

Completed a 3 month internship on Humanoid Robot Learning by Demonstration using Gaussian Mixture

Models.

Intern at Texas Instruments, India May 2008 - July 2008

Completed a 3 month internship focusing on Data Compression in Video Encoding Techniques.

Research Trainee at WAran Research FoundaTion, India June 2006 - May 2008

Completed a 2 year Research Training program with specialization in Signal Processing.

Programming - C, C++, Java, Python, Matlab

Computing Skills

Softwares - ROS, OpenCV, Tekkotsu, Fawkes

Assembly Language - AVR Microcontroller, 8051, 8086, 8085

Operating Systems - Unix and Windows

Masters Thesis - Analysis of the e ect of human interactions on Reinforcement Learning algorithms, with

Ongoing Projects

focus on methods like Learning by Demonstration and Apprenticeship Learning. Advisor - Prof. Michael

Littman

Robot Learning by Demonstration using GMM s (KBSG Lab, RWTH University) July 2009

Completed

A behavior acquisition model was developed for the Nao s using Gaussian Regression. After generalizing

Projects

the kinesthetic demonstrations, the robot was used to imitate constrained reaching gestures. Advisor - Prof.

Gerhard Lakemeyer

Best Narration Award - Introduction to Reinforcement Learning (RL3 Lab) April 2009

The Lego Mindstorms was programmed to learn real-time in a deterministic environment and to build a

model of the world using concepts of Graph Search and Dynamic Programming. A video tutorial was

submitted to IJCAI 2009. Advisor - Prof. Michael Littman

Autonomous Object Recognition using Corner Detection (Rutgers) December 2008

Implemented using Corner Descriptors and Geometric Point Matching methods. The advantage of the

system was the reduced number of descriptor points as compared to the SIFT algorithm. Advisor - Prof.

Lawrence Rabiner

Parallel Particle Swarm Optimization (Rutgers) December 2008

Parallel implementation of the PSO algorithm using MPI. The aim is perform a comparative analysis with

the sequential algorithm and to test its application for Multi-Agent Systems. Advisor - Prof. Manish

Parashar

Mobile Video Reference Data Compression (TI) July 2008

Developed transform-based techniques using C to compress the reference data acquired from videos captured

using mobile phones. This technique was implemented in the H.264 standard. Advisor - Mr. Ajit Gupte

Thomas J. Walsh, Kaushik Subramanian, Michael L. Littman, Carlos Diuk: Generalizing Apprenticeship

Publications

Learning across Hypothesis Classes. To appear in ICML 2010, Haifa, Israel, June 2010.

Kaushik Subramanian: Task Space Behavior Learning for Humanoid Robots using Gaussian Mixture Regres-

sion. To appear in AAAI 2010, Atlanta, USA, July 2010.

AAMAS 2010 ALIHT Workshop in Toronto, Canada, May 2010

Workshops

Presentation on Generalizing Apprenticeship Learning across Hypothesis Classes.

Dhi Yantra 2008, Workshop on Supercomputing and Brain Modeling conducted by the WARFT, India.

Presentation on Higher Order Gabor Statistics for Speech and Image Signal Feature Extraction.

Ardent fan of Origami and Sudoku.

Extra-Curricular

Trained in singing and can also play the guitar and keyboard.

Activities

Active participation in Soccer Events.

Contact this candidate