YUTI KADAKIA
adi790@r.postjobfree.com +1-312-***-**** linkedin.com/in/yutidkadakia/
EDUCATION
Master of Science, Electrical and Computer Engineering Aug 2019 – May/June 2021 (Expected)
University of Illinois at Chicago GPA: 4.0
Bachelor of Engineering, Electronics and Communication Aug 2013 – May 2017
Visvesvaraya Technological University (VTU) GPA: 3.71
TECHNICAL SKILLS
Simulation Tools: ModelSim, Cadence Virtuoso, Intel Quartus Prime
Programming Languages: C, Python, Verilog, System Verilog, x86 Assembly Language (Basics), MATLAB
Operating Systems: Windows, Ubuntu
Concepts: VLSI Design, Digital System Design, Computer Architecture, Neural Networks (Deep Learning)
ONGOING THESIS
Energy Aware Network Operator Search (ENOS)
•Developing a hardware-oriented design for a reconfigurable Processing Element (PE) Core in a Deep Neural Network Accelerator inspired by the DARTS (Differential Architectural Search) setup
•This will allow the PE to select a suitable operator for the MAC unit in each layer of the Neural Network architecture with the goal to optimize performance and suit the energy requirements of the accelerator
•The current setup allows the user to choose from four operators (Typical, Multiplication-free and Binary Operators) and is tested for CIFAR-10 and CIFAR-100 for four Neural Network Architectures – MobileNetV2, SqueezeNet, ShuffleNet and VGG16
•The energy consumption of all these operators is calculated using a reconfigurable PE setup in Verilog/System Verilog with synthesis to be done on Cadence RC (learning TCL/Perl for the same)
PROJECTS
MIPS Single Cycle and Multi Cycle Processors
•Designing MIPS Single Cycle and Multi Cycle Processors in ModelSim and Quartus Prime using System Verilog
Implementation of Memory Scheduling Algorithms on the USIMM Simulator
Implemented Thread Cluster Memory Scheduling (TCM) and Parallelism Aware Batch Scheduling (PAR-BS) Algorithms using FR-FCFS as a base algorithm on the USIMM Simulator
Implementation of Spiking Neural Network (SNN) for Neuromorphic Computing and proposing suitable device(s) for its efficient processing
A different approach for firing a spiking neuron is implemented and presented a survey analysis of suitable devices for hardware implementation of SNNs
Multiplication and Accumulation (MAC) Datapath for Neural Networks
•Implementation of MAC Unit focusing on reduction in power consumption and lesser delay using Cadence Virtuoso
Developed an LPG Gas Leakage Detection and Prevention System
An accurate and sensitive LPG gas leakage detection system using a MQ-6 sensor which will alert the user using GSM Module and turn a buzzer on and the leakage is prevented by stopping the gas supply using a stepper motor and the AC mains using a relay.
EXPERIENCE
Software Engineer, Accenture Solutions Pvt. Ltd. Sep 2017 - Jun 2019 Worked on an Accenture Product CAS - mainly deals with the Account Plan, Payment Management and Promotion Planning sector of these industries.
Development of Trade Promotion Management related application as per the client requirement using VB ASP.NET.
Provided Application Support to various Consumer Goods Clients such as MARS, Minute Maid and Coca-Cola, P&G
ONLINE CERTIFICATIONS
•Neural Networks and Deep Learning - Coursera
•Python Data Structures - Coursera
•Introduction to FPGA Design for Embedded Systems - Coursera
•Introduction to Programming using MATLAB - Coursera
•Programming for Everybody (Getting Started with Python) - Coursera
•‘Embedded System using ARM7 and ZigBee’ workshop certification