SINDHU NARAYANASWAMY

addobg@r.postjobfree.com 469-***-**** 45022 Deepwood Ct, Shelby Township, MI 48317 linkedin.com/in/Sindhu-Narayanaswamy SUMMARY

Software engineer with experience in developing real-time Signal Processing algorithms; Optimized Kalman Filter and Deep Learning Models for Sensor Fusion, Object Recognition and Speech Enhancement applications; Result oriented, curious and have an innate desire to learn new things and work collaboratively to find solutions and celebrate success EDUCATION

Master of Science – Electrical Engineering GPA: 3.80/4 May 2020 University of Texas at Dallas, Richardson, TX

Coursework: Digital Signal Processing, Probabilities, Random Variables and Statistics, Linear systems, Pattern Recognition, Machine learning, Speech and Speaker Recognition, Optimal Estimation and Kalman Filter, Applied Digital Signal Processing Bachelor of Engineering – Electrical and Electronics Visvesvaraya Technological University, KA, India

GPA: 3.85/4 May 2016

PROFESSIONAL EXPERIENCE

Software Engineer, Bosch, Bangalore, India Sep 2016 – Jul 2018

• Implemented State estimation (Kalman Filter), sensor fusion algorithms, feedback control system and evaluated Vehicle traces for Electronic Stability Controller functionalities based on customer requirements, while ensuring functional safety.

• Developed AUTOSAR embedded software for Braking System ECU following V-model of ASPICE, from requirement analysis using IBM Doors to unit test, validation and verification to achieve ASPICE compliance.

• Developed Diagnostic Software for Application SW, and Basic SW (CAN, DCOM) fault monitoring and ensured software quality with unit test, component test, MISRA, defect fix, errata fix, SIL, HIL using LABCAR, Vector tools and debuggers. Embedded Software Engineer Intern, ETAS Inc., Ann Arbor, MI May 2019 – Aug 2019

• Developed Application layer for Body Control Module to realize automatic headlights and wiper control; Simulated the state flow diagram and control system using model-based tool ASCET to ensure algorithm robustness.

• Integrated AUTOSAR architecture, Configured Basic Software (BSW) for Infineon Aurix-TC277 ECU using ISOLAR- A (AUTOSAR authoring tool) and improved development velocity by optimizing AUTOSAR workflow.

• Integrated and Tested ECU software using ISOLAR-EVE (virtual ECU) tool leveraging the AUTOSAR standard, thus enhancing software quality at an early point in the development process. Research Assistant, University of Texas at Dallas, Richardson, TX Jan 2019 – May 2019

• Implemented Feature extraction algorithms such as Image segmentation, adaptive filtering, spectral, local binary pattern, Haralick features and linear predictor models to extract features from EEG signals to improve seizure prediction.

• Developed Feature selections model to select best features from EEG which reduced the hardware complexity and achieved 5% increase in the accuracy. Also implemented optimal Machine learning algorithms to predict seizures. Software Engineer Intern, Indian Space Research Organization, Bangalore, India Jan 2016 – Apr 2016

• Designed a system-level Simulink model for 16-Point FFT to validate proposed DSP architecture. Verified 4096 Point FFT algorithm accuracy using MATLAB code, Programmed Radix-22 SDF(DSP) algorithm in VHDL and targeted into Xilinx Virtex-4 FPGA, which optimized HW utilization to 81% on FPGA, and improved OFDM system performance. PROJECTS

Adaptive Cruise Control (ACC) and Lane Centering for robot vehicle Apr 2020 – Present

• Developed Lateral and Longitudinal control of robot vehicle in C with PID controller and Ultrasonic Sensor on ATmega328 microcontroller. Calibrated the sensor and designed Kalman filter to determine the best State Estimation. Real-Time Implementation of DNN on Android for Detection of Diabetic Retinopathy Mar 2020 – May 2020

• Developed Deep Neural Network model using transfer learning to detect Diabetic Retinopathy and achieved test accuracy of 81%. Also optimized the implementation of real-time Image processing on Android. Self-Supervised Deep Learning-Based Speech Enhancement Oct 2019 – Dec 2019

• Trained Convolutional Neural Network model using MATLAB and Python for Speech enhancement in environments with machinery & babble noise without any clean speech (Self-supervised model).

• Optimized to improve Output SNR by 37.5%reduced software latency to 56.4 micro-seconds. Speaker Identification on Fearless Steps: Apollo 11 NASA Corpus Nov 2019 – Dec 2019

• Developed a speaker recognition model based on traditional Gaussian Mixture Model and Deep Learning Model using Kaldi tool (C++, python and shell scripting) to identify speakers, and increased Top-5 Accuracy by 10%.

• Implemented pre-trained i-vector and x-vector model for each speaker using NASA Apollo 11 Corpus Dataset. Pedestrian recognition using Deep Learning Feb 2019 – Mar 2019

• Developed (Faster R-CNN) Deep-Learning based model using Keras framework for pedestrian recognition which was pre-trained on Nvidia GPU with KITTI vision Benchmark dataset to improve the software latency and accuracy. TECHNICAL SKILLS

Programming/Engineering tools: C, C++, Python, Matlab, Simulink, LabView, Android Studio, Keras, Kaldi, OpenCV,SVN Sensors: Ultrasonic, IR, Camera, IMU, wheel speed sensor, steering angle sensor, RADAR, GPS Automotive Skills: ABS, TCS, ESP, ACC, VSE, Sensor Signal Processing Component, Sensor Fusion, Perception Automotive tools: AUTOSAR, ASPICE, ISO 26262 standards, ISO 14229, MISRA, Vector Canalyzer, ETAS tools Machine Learning: Adaptive, Wiener & Kalman Filter, MFCC, AR, MA & ARMA models, CNN, Regression, Classification

Contact this candidate