MOHAMED ELWAKDY

Contact:

at +*- (***) ***- ****

Email: ad1iha@r.postjobfree.com

Peoria, Illinois - USA

SUPPLIER PARTNER SUMMARY

I am a professional and technically talented data analyst with long exposure to the data manipulation, data cleansing, simulation, and visualization and specialized in Machine Learning, Artificial Intelligence, Signal and Image Processing, Data Mining & Data Analytics, with a solid background in Pattern Recognition, Action Recognition, Behavioral Recognition and Speech Recognition with a very good eye on the details. I can play a role in analyzing problems and come up with creative solutions for effective data management. I have extensive experience in Big Data Processing and building Algorithms/Predictive models from scratch.

• Outstanding commercial and Academic/Research experience in areas such as: clustering, k-means, and Support Vector Machines, Neural Networks, Decision Tree, Linear Regression and Random Forest.

• Experience in Jupyter and python libraries Numpy, Pandas, Scikit-Learn, MatPlotLib, requests and Flask.

• Statistical Analysis, and Hypothesis testing with familiarity with core statistical concepts such as the design of experiments, and confidence intervals and p-values.

• Experience in computer vision tasks (object detection and tracking, and classification).

• MATLAB, R and Python language skills, and hands-on experience with C++ and Tableau.

• RESTful Web services, Web Scraping (Data Scraping) and Amazon Web Services (AWS).

• Good knowledge of Deep Learning (CNN), Natural Language Processing (NLP).

• Good Knowledge in SQL and Snowflake

• Strong knowledge of data science / modeling workflow with experience working a combination of big data/ advanced data analytics.

PROJECTS

• How Humidity, Temperature and Wind Speed effect on increase the infection rates of COVID-19 in 10 Counties (population between 55,000 to 85,000) in New York State.

• Analyze a dataset which contains information about taxi rides in NYC (quite large dataset).

• Prediction of housing price using Python to get an idea of how the price changes with respect to each aspect with data cleansing using housing dataset.

• Prediction of the sale price of bulldozers sold at auctions using Python. The Random Forest Regressor has been used for the classification. I received 91.12 % classification accuracy.

• Prediction of forecast product sales using Python to see how machine learning could better ensure they please customers by having just enough of the right products at the right time. The Random Forest Regressor has been used for the classification.

• Prediction of matching the addresses in the dataset “address_matching_data” with a classification accuracy of 98% using Python. XGBClassifier is used for classification. TECHNICAL EXPERIENCE

SENIOR DATA SCIENTIST (employee) January 2022 - Current Caterpillar Illinois – USA

• Implemented two Machine Learning models that determine aftercooler, fuel injector, and air system leaks on the engine systems of 500+ Caterpillar large mining trucks.

• Algorithm to identify High exhaust temperature events for CAT Medium Duty trucks and tractors. MOHAMED ELWAKDY

• Algorithms to detect low crankcase pressure events on D3500 Units for a list of prefixes/serial numbers. This included finding out the start and end date for each trigger/event, duration in seconds for each trigger/event, number of points for each trigger/event, number of triggers/events, the frequency of each duration time and percentage of Events/triggers to the total number of triggers/events in addition to the total number of assets under each trap condition.

• Comparing 1 Hz algorithms before and after data aggregation using stats functions to increase the value and selling opportunity.

• Algorithms to determine anomalies in Left and Right Banks for the following systems prior to bottom end failures: Crankcase Pressure, Engine Speed, Engine Load, Engine Oil Pressure, Engine Oil Temperature, Engine Oil Filter Differential Pressure, Engine Coolant Temperature, Engine Exhaust Temperature.

• Counted the number of hot shutdowns for several assets towards determining if engine design changes are necessary.

• Created and implemented an algorithm to determine the mean oil level and oil volume over each five- service hour intervals and change in oil level for each twenty service hour intervals by detecting the areas where the oil level is stable and unstable. Also investigated how long it takes for oil to become stable and how long after ground speed is zero oil level becomes stable.

• Filtered the data from all codes (event and diagnostic codes) that happen due to sensor faults based on Auto-created threshold.

• Data study to find whether there is a correlation between the increase of iron rate, copper rate, silicon rate, soot rate and lead rate in engine oil and many events codes related to the coolant temperature, exhaust temperature, oil level, oil temperature and oil pressure in addition to oil change interval and the total number of engines running hours.

• Data study to determine whether the silicon rate, soot rate, copper rate, oxidation, nitration and sulfonation are running high before the oil cooler failures. DATA ANALYST (employee) January 2021 – January 2022 Super-Pufft Florida – USA

• Data Collection from different machines (packers and case packers).

• Counted the number of good and bad bags based on datetime variable in the dataset to evaluate the operators’ performance over each shift.

• Determined the time spent troubleshooting ISHIDA packers and case Packers during each shift.

• Counted the errors resulting from both machine operators and machines themselves. This will assist the owners in identifying the individual strengths and weaknesses of each operator, enabling them to provide focused training to address any identified weaknesses.

• Created a weekly performance evaluation for each operator, enabling them to evaluate their own performance over the week. This will help the operation managers and owners in assessing the performance of each operator on a weekly basis.

• Assign individual performance targets to operators for each shift, based on the specific capabilities of each machine.

DATA ANALYST INTERN/ TEAM LEADER September 2020 – June 2021 SeedStages California – USA

• Built predictive models using machine learning algorithms.

• Data cleansing on the sales dataset using Pandas and Numpy packages.

• Extracted the information (Data cleansing) from messages and opportunities datasets to analyze job seekers’ messages and job description details, comparing their skills/qualifications to employers’ requirements.

• Extracted the information (Data cleansing) from “Record Screen Views” dataset to identify user behaviors, including the total count of screens viewed by users (students/companies) and a list MOHAMED ELWAKDY

of datetime activities such as datetime-in and datetime-out for each user based on the IP Address.

• Calculate the duration that each user spends on each screen in the application by splitting this time into multiple sessions if the user leaves the application and returns after more than 5 minutes.

• Assigned tasks to graduate students, which involve building models utilizing machine learning algorithms.

TEACHING ASSISTANT (employee) January 2020 – May 2020 Clarkson University New York- USA

• Teaching Econometrics.

MACHINE LEARNING TECHNICAL CONSULTANT (consultant) April 2017 – July 2018 StatsLab Business Intelligence New Zealand

• Provided training to recent graduates in how to build machine learning models.

• Transformed data into actionable insights and effectively engaging with StatsLab clients.

• Utilized the K-NN algorithm (with an accuracy of 98%) and Linear Regression (using R/Python) for predicting cancer diseases. The model facilitates early detection of abnormal lumps or masses in breast tissue.

• Utilized the Decision Tree Algorithm (using R) to predict default risk (with an accuracy of 74%). The credit approval model aims to identify factors that increase the likelihood of loan default, considering information about the applicant and whether the loan eventually defaulted.

• Used Linear Regression (using R) to predict medical expenses for the insured population (with an accuracy of 75%). The model estimates average medical care expenses for different population segments.

• Created education materials for paid workshops utilizing diverse datasets and machine learning algorithms.

UNIVERSITY LECTURER AND RESEARCHER (employee) November 2004 – June 2015 Faculty of Industrial Education – Helwan University Egypt

• Taught a range of subjects, including Electrical Circuits, Electronic Circuits, Computer Networking, Data Communications, Systems and Networking Industrial Electronics and Television and Video Systems.

• Worked as a supervisor for multiple undergraduate students’ graduation projects, utilizing Python.

• Worked as a researcher in the field of Signal Processing, particularly in Speech Recognition and Pattern Recognition.

• Built ‘Developed Trajectories Classification Algorithm (DTCA)’ to accurately recognize on different moving objects with similar trajectories, achieving an accuracy of 98.83%.

• Utilized MATLAB to recognize individuals walking and running (with an accuracy 99.9%) based on the movement of their heads in two dimensions (X, Y). The Vicon Physical Action dataset was used for this research study.

RESEARCHER ASSISTANT (employee) February 2013 – January 2014 Department of Biophysical and Electronic Engineering

– ISIP40 Research Group – University of Genoa

Italy

• Worked as a researcher in the Object Tracking Project, focusing on objects detection and object tracking and Ships’ trajectories classification. This project received support from the Italian Navy.

• Utilized Computer Vision technology in MATLAB to detect multiple objects in the Region of Interest

(ROI) and extracted their trajectories over time.

MOHAMED ELWAKDY

• Used a Kalman filter to predict the location of the trackers within the ROI in each frame.

• Built the Trajectories Classification Algorithm (TCA) in MATLAB to accurately discriminate between the trajectories of different objects, specifically tanker ship and fishing boats, achieving an accuracy of 99.9%.

EDUCATION

Master Degree Data Analytics (GPA: 3.76) - Clarkson University – NY, USA - 2020. PhD Degree Electrical Engineering (ML – Pattern Recognition) - Ain Shams University – Cairo, Egypt – 2018.

Master Degree Electrical Engineering (ML – Speech Recognition) - Ain Shams University – Cairo, Egypt

– 2008.

Bachelor of Science Electrical Engineering - El Obour Academy for Engineering and Technology – Cairo, Egypt – 2001.

Associate Degree in Computer Maintenance - Northampton Community College – PA, USA – 2010. Specialized Diploma in Multimedia Production - Northampton Community College – PA, USA – 2010. SCHOLARSHIPS/AWARDS/GROUP MEMBERSHIP

CLARKSON UNIVERSITY SCHOLARSHIP

• Postgraduate Study (master’s in data Analytics)

FULBRIGHT ORGANIZATION SCHOLARSHIP

• Community College Initiative (CCI) program

UNIVERSITY OF GENOVA - ISIP40 research Group

• Research Grants

PUBLICATIONS

Published Paper "A Novel Spatial Algorithm for Similar and Non-Similar Trajectories Classification of Different Objects” in:

- FECS'15- The 2015 International Conference on Frontiers in Education: Computer Science and Computer Engineering, Las Vegas, USA, 2015.

Published Paper "A Novel Trajectories Classification Approach for Different Types of Ships Using a Polynomial Function and ANFIS” in:

- IPCV'15 - The 19th International Conference on Image Processing, Computer Vision, & Pattern Recognition, Las Vegas, USA, 2015.

Published paper “Speech recognition using a wavelet transform to establish fuzzy inference system through subtractive clustering and neural network (ANFIS)” in:

- Proceedings of the 12th WSEAS International Conference on Systems, Heraklion, Greece

(July 22-24 – 2008).

- International Journal of Circuits, Systems and Signal Processing Issue 4, volume 2, 2008 North Atlantic University (NAUN).

Contact this candidate