Summary
Overview
Work History
Education
Skills
Publications
Other Projects
Languages
References
Timeline
Generic

Arya Yazdani

San Diego,CA

Summary

Results-driven and highly motivated Software Engineer with 4 years of experience in Computer Vision and 10+ years of experience in software development. Adept at coding in Python, Machine Learning, and Object Detection, with a proven track record of hands-on experience in the implementation of object detection and classification for real-time safety measurement purposes. Seeking to leverage my skills and expertise in ML and specifically Computer Vision to contribute effectively to development and team leadership, driving innovations in technology, cost reduction, and improving efficiency and reliability.

Overview

9
9
years of professional experience

Work History

Senior Software/Computer Vision Engineer

ACCELERA
01.2023 - Current
  • Developed, tested, debugged, implemented, and maintained deep learning and computer vision algorithms on hardware integrated into vehicles to extract valuable information about the road using vision data
  • Developed communication protocols between computer vision hardware and the vehicle supervisory controller to address road hazards and implement safety measures
  • Designed new algorithms and conducted testing for commercialization purposes
  • Increased efficiency and quality of managing release and production software by supervising the use of DevOps tools such as Git, and assisted other engineers in resolving issues
  • Utilized tools including Python/Jupyter Notebook, TensorFlow, Linux, bash scripts, SSH, OpenCV, Convolutional Neural Network (CNN), Deep Learning, Confusion matrix, Precision-Recall curve, Neural Networks, MATLAB/Simulink, Git/GitHub, XML, Polarion.

Computer Vision Research Engineer

SAN DIEGO STATE UNIVERSITY
09.2020 - Current
  • Lead engineer on the project titled "Real-time vehicle and pedestrian object detection and classification on the Coral EdgeTPU Board for Surrogate Safety Measurements." Captured 17,000 images from a live camera stream using OpenCV for real-time object detection and classification on a low-cost Edge TPU board (less than $150)
  • Developed, optimized, and debugged Python scripts to organize and categorize images before and after annotations were done
  • Trained Convolutional Neural Networks MobileNet V1 and V2 on a remote Ubuntu machine through SSH using eight Nvidia Tesla V100 PCI-e GPUs, Python scripts, and Jupyter Notebook
  • Compiled a light version of the trained model (TFLite) for the Google EdgeTPU board for real-time object detection and classification
  • Implemented the TFLite trained network in a service script, enabling real-time detection and classification of pedestrians and different vehicles
  • Troubleshooted and updated Python scripts to be compatible with TensorFlow 2 and the specific case in this project
  • Analyzed results in terms of neural network performance metrics such as Precision-Recall (PR) curve, mean Average Precision (mAP) at different Intersection over Union (IoU) values, and real-time inference speed
  • Tools: Python/Jupyter Notebook, TensorFlow, Linux, bash scripts, SSH, OpenCV, Convolutional Neural Network (CNN), Deep Learning, Confusion matrix, Precision-Recall curve, Neural Networks.

Senior Software/Controls Engineer

MERITOR
08.2018 - 12.2022
  • Led the development of software for powertrain and accessories in heavy-duty electric vehicles, with a focus on achieving the production of zero-emission powertrains
  • Played a leading role in multiple income-generating projects for the company as a lead software engineer
  • Utilized Python, C, C++, AUTOSAR APIs, and MATLAB Simulink to develop software for vehicle controllers, implementing efficient algorithms
  • Tracked teamwork in daily meetings and improved processes based on progress feedback
  • Established processes for nearly 100% automated software development and tracked progress using lifecycle management tools like Polarion
  • Troubleshooted issues in released software and automation Python scripts by identifying the root cause and creating prompt action plans, including corresponding modifications
  • Implemented skills acquired from 8D Problem-Solving, MCE 7 Tools, and Statistics training to improve the software development process
  • Worked in a team environment and effectively communicated software-related concepts with peer engineers and non-technical stakeholders
  • Utilized tools including Python/Jupyter Notebook, MATLAB/Simulink, C/C++, Linux bash/Windows batch, AUTOSAR, Git/GitHub, XML, CAN communication, 8D Problem-Solving, Polarion.

Test Engineer

HONDA R&D AMERICAS
04.2017 - 07.2018
  • Experience as a test engineer involved in testing production powertrains for Honda and Acura vehicles
  • Made engineering judgments on vehicle powertrain performance using tools such as: Analyzed the vehicle cooling system by determining parameters in a P-diagram, including system input and response, system noise factors (part to part variation, degradation of equipment), control factors (pipe/hose and flow meter dimensions), and error states (malfunctioning flow meter/pressure transducer/thermocouple, leaks in the cooling system)
  • Applied electric analogies for different components in the engine cooling circuit to determine performance expectations and judge test results
  • Utilized Fault Tree Analysis (FTA) to predict and diagnose engine cooling system false measurements of system states such as flow, pressure, and temperature
  • Used tools including FTA analysis, P-Diagram, and Microsoft Office.

Research Engineer

MICHIGAN TECHNOLOGICAL UNIVERSITY
09.2014 - 04.2017
  • Lead research engineer working with the Ford Motor Company powertrain team to develop algorithms aimed at increasing efficiency and reducing costs
  • Led algorithm generation in a project titled: "Air Estimation in VCT Engine Utilizing In-Cylinder Pressure Sensors" sponsored by Ford Motor Company
  • Utilized theory to generate estimator codes and performed uncertainty analysis in EES and MATLAB software
  • Calibrated the algorithm using test data from experiments on the experimental setup
  • Validated and verified the estimation algorithm
  • Used tools including MATLAB/Simulink.

Education

Master of Science - Electrical and Computer Engineering

San Diego State University
08.2021

Master of Science - Mechanical Engineering

Michigan Technological University
08.2016

Bachelor of Science - Applied Mathematics

K.N. Toosi University of Technology
09.2009

Skills

  • Python
  • C
  • C
  • C#
  • Linux bash
  • Windows batch
  • TensorFlow
  • Git
  • OpenCV, Image formation, capturing, and processing pipeline
  • XML
  • Vim editor
  • Jupyter Notebook
  • SSH
  • MATLAB/Simulink
  • AUTOSAR
  • CAN Communication
  • Deep Learning
  • Machine Learning
  • Excellent communication and leading skills
  • Problem-solving skills
  • Great team player

Publications

1. Real-time optimal control of power management in a fuel cell hybrid electric vehicle: A comparative analysis, SAE International Journal of Alternative Powertrains, 2018. 

2. Air charge and residual gas fraction estimation for a spark-ignition engine using in-cylinder pressure, SAE Technical Paper, 2017. 

3. A comparative analysis for optimal control of power split in a fuel cell hybrid electric vehicle, SAE Technical Paper, 2016. 

4. Air charge estimation for an SI engine using in-cylinder pressure sensor, Michigan Technological University, 2016. 

5. Modeling, performance simulation and controller design for a hybrid fuel cell electric vehicle, Journal of the Brazilian Society of Mechanical Sciences and Engineering, 2015. 

6. Dynamic modelling and simulation of a polymer electrolyte membrane fuel cell used in vehicle considering heat transfer effects, Journal of Renewable and Sustainable Energy, 2012. 

7. Modeling and simulation of a PEM fuel cell (PEMFC) used in vehicles, SAE Technical Paper, 2012.

Other Projects

Dynamic Control of a PEM Fuel Cell Hybrid Vehicle, IFCO Company

  • Simulated and validated a comprehensive model of a fuel cell hybrid electric vehicle.
  • Utilized a V-diagram to model and verify the vehicle by defining subsystems such as PEM fuel cell, Longitudinal Vehicle Dynamics (LVD), battery, and electric motor in Simulink.
  • Validated the fuel cell model using test data from an experimental setup found in the literature.
  • Conducted Model-in-the-Loop (MIL) testing on the simulated vehicle powertrain to verify the performance of designed conventional controllers.

Optimal Control of a Fuel Cell Hybrid Electric (FCHEV) Vehicle using DP and MPC Control Strategies

  • Conducted MIL testing to verify two optimal control strategy designs, dynamic programming, and model predictive techniques, to control power split between the fuel cell and battery in a fuel cell hybrid vehicle for a specific driving cycle.
  • Investigated the trade-off between the global optimal solution and real-time implementation.

Languages

English
Full Professional
Persian
Native/ Bilingual
Spanish
Limited

References

Available upon request.

Timeline

Senior Software/Computer Vision Engineer

ACCELERA
01.2023 - Current

Computer Vision Research Engineer

SAN DIEGO STATE UNIVERSITY
09.2020 - Current

Senior Software/Controls Engineer

MERITOR
08.2018 - 12.2022

Test Engineer

HONDA R&D AMERICAS
04.2017 - 07.2018

Research Engineer

MICHIGAN TECHNOLOGICAL UNIVERSITY
09.2014 - 04.2017

Master of Science - Electrical and Computer Engineering

San Diego State University

Master of Science - Mechanical Engineering

Michigan Technological University

Bachelor of Science - Applied Mathematics

K.N. Toosi University of Technology

Similar Profiles

CREATE PROFILE
Arya Yazdani