Educational Background
University of California San Diego (UCSD)
Ph.D. in Electrical Engineering (Communication Theory and Systems), GPA: 4.0 09/2009–06/2013
Dissertation topic: Active Learning and Hypothesis Testing
Advisor: Professor Tara Javidi
M.Sc. in Electrical Engineering (Communication Theory and Systems), GPA: 4.0 09/2007–09/2009
Sharif University of Technology (SUT)
B.Sc. in Electrical Engineering, GPA: 18.79/20 09/2003–07/2007
Honors/Awards
Ph.D. Fellowship 09/2007–06/2008 Department of Electrical and Computer Engineering, University of California San Diego
Graduated with rank 3 among 175 students 07/2007 Department of Electrical Engineering, Sharif University of Technology
Ranked 1st among 500,000 participants in Iran’s private universities entrance exam 07/2003
Ranked 8th among 600,000 participants in Iran’s nationwide university entrance exam 06/2003
Work Experience
Amazon
Senior Machine Learning Scientist at Amazon Music 10/2021-Present
Machine Learning Scientist at Amazon Music 02/2021-09/2021
Machine Learning Scientist at Amazon Fresh and Physical Stores 05/2019–01/2021
Qualcomm Technologies Inc. 05/2017–05/2019
Staff Systems Engineer in the Corporate Research and Development Department.
Member of Qualcomm AI Research team working on reinforcement learning and deep learning with focus on autonomous driving.
Qualcomm Technologies Inc. 06/2013–05/2017
Senior Systems Engineer in the Corporate Research and Development Department.
Member of Qualcomm research systems engineering team working on the design, analysis, prototyping, and development of LTE-U (LTE advanced in unlicensed spectrum). Developed coexistence algorithms for LTE-U including carrier sense adaptive transmission (CSAT), channel selection, and medium utilization (MU) estimation for efficient spectrum sharing with other technologies deployed in the unlicensed band.
ASSIA Inc. 06/2011–09/2011
Summer internship on design of diagnostic tools for DSL transmission lines.
Developed an algorithm to detect a missing microfilter (main cause of line instability) by processing and analyzing data collected from a single-ended loop test (SELT). Conducted lab tests to measure the performance of the algorithm. Designed a circuit board to model telephones in off-hook and on-hook positions which significantly expedited lab experiments and enabled remote testing.
Micromodje Industries 06/2006–09/2006
Internship as part of the bachelor’s degree requirements.
Studied different types of wireless communication networks with a focus on trunked radio systems. Compared the features of various trunked radio systems such as Logic Trunked Radio (LTR) and Terrestrial Trunked Radio (TETRA).
Research Experience
Active hypothesis testing and optimal experimental design 09/2009–06/2013
Studied the problem of identifying the true hypothesis in a speedy manner and with low probability of error by sequentially selecting actions among the set of available ones. Introduced Extrinsic Jensen–Shannon (EJS) divergence as a new measure to quantify the information gained by different actions, and proposed a heuristic based on greedy maximization of EJS. The proposed algorithm, under some technical conditions, detects the true hypothesis at the maximum possible rate with the highest reliability, and has applications in active learning, communication with feedback, and noisy dynamic search. The results have been published in major conference proceedings and journals including the prestigious Annals of Statistics.
100% throughput routing in constrained queueing networks 01/2009–09/2009
Combined ideas from opportunistic shortest path and backpressure routing to arrive at a routing algorithm with good delay performance. Derived a class of continuous, differentiable, and piece-wise quadratic Lyapunov functions which were used systematically to establish the throughput optimality of the proposed routing algorithm.
Adaptive routing in ad-hoc networks via reinforcement learning 6/2008–01/2009
Studied the problem of packet delivery in wireless ad-hoc networks in the absence of knowledge about the underlying stochastic model of the network. Reinforcement learning techniques were applied to design an adaptive routing algorithm that achieves the optimal performance.
Professional Services
Session Chair, ITA 2013
Editorial Services
IEEE Transactions on Wireless Communications
IEEE Transactions on Information Theory
IEEE Journal of Selected Topics in Signal Processing
International Symposium on Information Theory
IEEE Information Theory Workshop