Behnoush Rezaeianjouybari Email: adjce0@r.postjobfree.com

https://www.linkedin.com/behnoush-rezaeianjouybari Mobile: +1-573-***-**** Authorized to work in US without sponsorship.

Education

Ph.D. in Mechanical Engineering (Minor in Statistics) Jan. 2016 - Mar. 2021 (expected) University of Missouri Columbia, MO

M.Sc. in Mechanical Engineering Dec. 2011

Sharif University of Technology Tehran, IR

Summary of Qualifications

5+ years of extensive experience on working with machine and deep learning tools.

Skilled in advanced machine learning technologies including transfer learning and domain adaptation.

Authored 10+ journal and conference articles, with 250+ citations.

Proficient in scripting languages, cloud computing technologies and mechanical engineering commercial softwares.

5+ years hands-on experience in sensors and data acquisition (DAQ) instrumentation.

Skilled in advanced signal processing techniques in both time and frequency domains for extracting fault-related information from diagnosis sensory signals including vibration, acoustic and temperature.

Taught more than 400 students in graduate and undergraduate levels. Areas of Expertise

Transfer learning and domain adaptation Machine learning and deep learning State-of-the-art cloud computing technologies Script programming Advanced statistical analysis Data Visualisation Dynamic modeling of mechanical systems Design of experiments (DoE)

Signal processing Failure mode and effect analysis Condition monitoring of engineered systems Recent Working Experience

University of Missouri, Intelligent Manufacturing Laboratory Columbia, MO Research Assistant Jan. 2016 - Present

Developed a deep learning-based multi-source domain adaptation framework to transfer knowledge from various domains to one unlabeled domain -The proposed model uses task-specific decision boundaries based on sliced Wasserstein discrepancy metric and optimal transport theory (OT) to align feature distributions of various domains 30% accuracy improvement in machinery fault detection over non-cross domain alternative deep models (Boeing).

Implemented a generative convolutional neural network with modified adversarial cost function for cross-domain fault diagnosis of rolling element bearings based on multi-sensory heterogeneous data - 98% average accuracy under varying operating loads.

Designed and programmed a multi-sensory fusion-based model of fault diagnosis for centrifugal pump based on vibration, acoustic, temperature and pressure measurements using Support vector machines

(SVM), random forest and Dempster–Shafer evidence theory - 28% accuracy improvement over alternative non-fusion models (Boeing).

Constructed an encoder-decode-based Long-Short-Term-Memory (LSTM) recurrent network for Remaining useful life estimation of rolling bearings using pre-trained AlexNet architecture and vibration image inputs.

Trained and deployed a cloud-based condition monitoring framework by running python algorithms on Amazon EC2 instance using S3 buckets.

University of Missouri Columbia, MO

Teaching Assistant Aug 2016 - Present

Co-instructed Senior Capstone course for three semesters: Advising inter-disciplinary capstone projects, teaching engineering software and multi-body dynamic modeling.

Co-instructed Manufacturing design graduate course for two semesters: teaching quality engineering tools, Taguchi’s robust design method, FMEA, DoE, and design for additive manufacturing.

Instructed LabView programming lab for summer semester: fundamentals, programming structures, creating subvi’s, using local and global variables, instrument control and DAQmx API programming .

Instructed SolidWorks lab for two semesters: sketching, part design, drawing and assembly, motion study, and SimulationXpress.

Automotive Industrial Consulting Engineers, SAIPA group Tehran, IR Research Engineer Feb. 2013 - Jan. 2015

Investigated automation solutions for factory production lines improvements: working in a team to improve the automation levels of car production assembly lines, maintaining and troubleshooting the automation machines, design the control units, programming Omron and Siemens PLC - 153% production improvement.

Conducted reliability assessment and design validation through analytic methods including FEA, tolerance analysis, kinematic study and accelerated durability testing. Center of Excellence in Design, robotics and Automation Tehran, IR Research Assistant Nov. 2009 - Dec. 2011

Designed, trained and deployed an optimal control framework for trajectory planning of a redundant parallel industrial robot arm.

Adopted artificial neural network and evolutionary algorithms including particle swarm optimization and genetic algorithm for manipulability optimization of industrial robot arms. Selected Technical Skills

Statistics: Hypothesis testing, probability distributions, design of experiments, Bayesian statistics Programming: Python (4+ years), R(4+ years), SAS, SQL, MATLAB, OpenBUGS Machine Learning: Random forest, Gradient boosting, SVM, PCA, CNN, RNN, GAN, Auto encoder Deep Learning Frameworks: PyTorch, TensorFlow

Data Visualization: Matplotlib, Tableau, Grafana

Big Data & Cloud: AWS, IoT

Engineering: SolidWorks, Catia, MSC Adams, Ansys, AutoCad, PLC programming Honors and Awards

Accomplished Machine Learning, Stanford online course by Andrew Ng (Jun. 2019) [Certificate]

Accomplished Deep Learning Specialization, Deeplearning.ai online course by Andrew Ng (Jul. 2018) [Certificate]

Ranked among the top 0.1% of 650,000 participants of the National University Entrance Exam (2004).

Contact this candidate