1. Rajas Mhaskar
LinkedIn: RajasMhaskar Email: rajas.mhaskar@gmail.com Phone: 424-527-9680
Education
Master of Science, Computer Science Sept’19 – Dec’20
University of California, Los Angeles | GPA : 4.0 / 4.0
Bachelors & Masters of Technology, Engineering Design in Industrial Electronics July’12 –May’17
Indian Institute of Technology, Kharagpur | GPA: 8.41 / 10
Work Experience
Cadence Design Systems ,Tensilica IP, Pune June’17 –Sept’19
Design Engineer II, June ’17 – June’19 | Lead Design Engineer, July’19 – Sept’19
Vision Software: Developed optimized vision software for Tensilica Vision DSPs
o Monocular Visual SLAM: Reference C++ and DSP optimized code for modules such as Key point Detection,
Pose Estimation, Bundle Adjustment, Loop Closure and Bag of Words.
o Stereo Global Block Matching: Performance and memory optimization of the local cost computation modules.
o Alexnet: Performance optimization of the Alexnet network for Vision P6 processor
DSP Architecture: Worked on VLIW and SIMD Vision DSP architectures
o Vision Benchmarks: Analyzing the benchmark performance deviations due to pipeline changes in different
families of processors.
o Core Architecture: Experimenting with Instruction Set Architectures (ISA) by writing TIE to improve
benchmark performance cycles.
o Power Analysis: Analysed unit level power reports for Vision P6 and Vision Q6 DSP cores and improved the
core design to reduce power consumption for some benchmarks.
o ISS vs RTL: Debugged cycle performance deviations between instruction set simulators& RTL.
Cadence Design Systems ,Tensilica IP, Pune (Design Engineering Intern) Jan’16- June’16
Introduced MISRA-C embedded coding guidelines in the Audio DSP group.
Optimized multiple modules of the OPUS 1.1.2 audio codec for the HiFi3 audio DSP core.
Developed a Python based FFT tool for comparing two audio signals in frequency domain
Developed a Perl based systematic corruption technique for corrupting audio frame headers.
National Instruments India , Bangalore (Applications Engineering Intern) May ’15 – July ’15
Developed LabVIEW data acquisition modules for FPGA based NI myRIO.
Projects
Fault Detection & Diagnosis of Digitally Controlled Analog Circuits atPost-Silicon Level (Thesis Project, IIT)
Generated a fault table for mixed signal circuits by adding formal assertions, injecting faults and simulating on
Cadence Virtuoso.
Developed algorithms and implemented a C++ tool for producing optimal set of test vectors which can diagnose
and detect the faults in the fault list given the fault table.
Software Engineering Lab Term Project- Online Sales Portal
Designed (UML, SRS) and developed in Java an online portal for multiple buyers and sellers
Case Study for Acceleration of Graph Convolutional Networks on CPU and GPU
Achieved upto 100X gain on an Intel Xeon Gold CPU using AVX512 intrinsics and multi-threading using OpenMP.
Achieved a 200X gain on a Tesla P4 GPU by developing an optimized CUDA kernel for feature convolution
Term Paper – Improved TLB coverage by Exploiting Contiguous Memory Allocations
Analyzed & performed a comparative study of hardware and software (OS) based techniques for coalescing multiple
TLB page table entries for improving the TLB coverage as a part of the Advanced Computer Architecture course.
Publications/Presentations:
Co-author for the presentation titled ‘Fundamentals of Monocular SLAM’ presented at the Embedded Vision
Summit 2019 under the Fundamentals track in May 2019. (EVS 2019)
Delivered a Cadence Brown Bag presentation on Visual SLAM in July 2019.
Academic Achievements:
Graduated with Branch Rank 1 in the Engineering Design dual degree course, IIT Kharagpur.
Cleared the Indian National Mathematics Olympiad (INMO) in 2011 and was among the top 30 students selected
from India to attend the International Mathematics Olympiad Training Camp.
Honourable Mention at ACM Online Regional Contest in 2016 and 2017
Software Skills
Debuggers: GDB , Visual Studio. Programming Languages: C, C++, Java, Python, Perl
Softwares: Netbeans IDE. Scilab, Matlab, Simulink Parallel Programming Frameworks: CUDA, OpenMP
Teaching & Other Activities
Teaching Assistant for CS-M182 Systems Biomodelling and Simulation Basics, Fall 2019 at UCLA.
Delivered a lecture on Euclidean Geometry to high school students. (PreRMO camp IIT KGP)
Cleared CFA (Chartered Financial Analyst) Level 1 in December 2017