Ali Ebrahimi

Mobile: 613-***-**** Email: ac82wa@r.postjobfree.com linkedin.com/in/aliebrahimii

Current Technical Proficiencies

Programming Languages Python, R, MATLAB

Cloud Tech Microsoft Azure, Google Cloud Platform (GCP), Amazon Web Services (AWS), SOSCIP Databases SQL, MongoDB, ETL

Python Libraries NumPy, SciPy, SparkML, Pandas, Statsmodels, Matplotlib, Seaborn, Plotly, Scikit-Learn, TensorFlow, Keras R Libraries Dplyr, Ggplot2, Shiny, Plotly, Lubridate, e1071, TensorFlow, Keras, tseries, riskR, PerformanceAnalytics Tools Google Analytics, Hadoop, Spark, Hive, Knime, Tableau, Power BI, SAS, SPSS Operating Systems & VMs Linux (Cloudera), Windows, Mac OS, VMware and Oracle Virtual Box Econometrics Packages Stata, Rats, EViews, Dynare, Microfit, GAMS Document Preparation LATEX, Microsoft Office

Technical Skills Machine Learning, Data Mining, Statistical Modelling, Quantitative Analysis, Deep Learning, Optimization Soft Skills Teamwork, Flexibility, Communication, Decision Making, Adaptability, Time Management, Creativity Summary

As an experienced Data Scientist with six years of experience, and a strong background in economics. I have a good knowledge of data, especially the data generation process (DGP), Statistical Modelling, Time-Series Analysis and Machine Learning (Deep Learning, Data Mining, Evolutionary Algorithms, Pattern Recognition) and ETL. I also have valuable experience in agile methodology because of working as an intern in a well-known international company like Unilever. Having a good mixture of a strong background in economics, databases and programming languages (Python, R and Matlab), I did all of my previous carriers with a significant positive impact on their businesses which are mentioned below. Professional Experiences

Unilever Data Scientist Intern

April 2018 – January 2019 Toronto, Canada

Worked on: Trade Promotion Forecasting and Optimization. Develop and evaluate models that predict the impact of Unilever promotions on category and product share across different retailers (Market Share) and optimizing ROI. Concisely, I was working for optimizing Lift and ROI with Proof of Concepts since their dataset was huge; as a result we can predict the path of money, which is spent on promotions.

Visualizing datasets using Power BI and Tableau to find out trends and variable behaviours, and relations between variables.

Cleaning, Merging, and Filling the Gaps on six different datasets from different databases for five retailers through mapping datasets to connect promotions, the point of sale, loyalty, demographical and weather dataset to each other big dataset by ETL method, using SQL, MongoDB, and R.

Increasing the accuracy of our results and the team's decision making, by adding demographical, social economic and weather data of each location to the final dataset. In order to do this, I used Google API to fill the gaps between datasets.

Having 87 variables, I to choose and decide which ones have more power to explain our dependent variables (Lift and ROI). Increase the performance of the model by feature engineering techniques, and some feature selection techniques methods like LASSO, PCA and Linear Regression to choose relevant variables, using R. At the end, simple linear regression had the best result.

I used SVR in R to test the explanatory power of selected from variable selection part and make a prediction to choose the best scenario since according to variable selection; I had four different scenarios which were made in our think tank by my suggestion.

Predicting the target variables (ROI and Lift) by training a deep learning method (LSTM), as a result, testing the model that provide incredible results. Using Python and R Libraries: Keras, TensorFlow, SkLearn

Developing a toolbox in R to forecast the effect of spending on promotions on Unilever’s KPIs like ROI and Lift. The Ottawa Hospital Data Scientist (Research Assistance) June 2018 – August 2018 Ottawa, Canada

Worked on: Classification for predicting the priority of ICU using in Python.

Cleaning and labelling the dataset for using in classification by using statistical analysis; it was not a regular cleaning process since our dataset was related to patients and we had many gaps and missing values, so I tried to find the best way to fill some of them, which were required to have. Using Python and SQL

Classifying the dataset using SVM, Naïve Bayes and Decision Tree in Python; In this step I should figure out that which patient needed to transfer to ICU since if the patient need to go to the ICU and doctor did not send him, it has the risk of the die for the patient. Conversely, it would waste the money, so I classified the dataset to make a decision easier for doctors, and we have improved it around 10%. Capital Market Development Fund Data Scientist and Financial Analyzer January 2016 – September 2017 Tehran, Iran

Predicting the sign of price with the proof of concepts using deep learning (RNN) because of variety in the shares, in this case, I imported various variables like technical indicators, time series parameters, index of the industry, news information (as a dummy variable) and the total index as the explanatory variables. As a result, I got around 75% accuracy to predict the sign for daily data, in Python

Optimizing portfolio by multi-objective optimization algorithms (NSGA-II, SPEA-2, PSO-2 and ICA-2) and Reinforcement Learning to find the optimized portfolio on the efficient frontier of value at risk (VaR and CVaR) in MATLAB and R

Risk analyzing for a current stock portfolio and backtesting with statistical models (GARCH, EWMA, HS and MA) in R

Reporting and writing testing market efficiency (calendar and non-calendar effects) and structural changes on the Tehran Stock Exchange Market; as a result, I increased our efficiency by 5% to achieve our goal as a market maker. It is presented for Chairman of the Securities and Exchange Organization of Iran, Board of Directors and Stockholders.

Clustering most of the stocks to figure out the relations between them to predict the market direction, before clustering, I used heat map for selecting shares. As a result, I had found some clusters which the shares on them moved nicely together, using R (Innovative Project)

Setting up an algorithm-trading system, I did not have time and proper infrastructure to run it completely, but in the testing phase, its feedback was significant. For example, we could sell blocks of shares without an effect on the market in two weeks in the retail market. CompuCo Inter S.A Data Scientist Consultant

October 2015 – December 2015 Tehran, Iran

Worked on: Consultant in two projects

Evaluating the Credit Score and Segmenting commercial customers. In this project, I planned to cluster commercial customers using clustering methods (SOM, SVM and K-Means), and I had suggested them to select the best algorithm and way of using costumer’s information to score their credits in Python.

Predicting the volume of deposits in the bank; in this case, I used models, which captures the effect of seasonality because of my test results like SARIMA for modelling and forecasting in Python and R; as a result, we had about 70% accuracy. Quants Group at University of Tehran (Startup) Data Scientist and Economic Adviser October 2012 – September 2015 Tehran, Iran

Using clustering and classification methods for designing and developing the customer segmentation software for the Export Development Bank of Iran. As a result, we had a 65% accuracy.

Predicting ATM cash for FANAP Holdings by deep learning methods (RNN) for in Python; as a result, we had 80% accuracy.

Studying and reporting quantitative the behaviour of Lake Urmia at Iran: Previous, Present and Simulation of Future, for Lake Urmia restoration program to save it from drying. It is presented for the Vice-President and Urmia lake restoration program’s members.

Developing a Freud Detection toolbox for the Customs Administration to prevent smuggling and speed up the clearance process using data mining and deep learning.

Researching in Recidivism to build a smart system for reducing the number of prisoners using data mining.

Optimizing the routs to place wireless routers in one neighbourhood to reduce the cost as a proof of concept, using Machine Learning for Tehran Urban Planning & Research Center

Reporting Economic Evaluation of Cloud Wi-Fi in Tehran as a part of building a smart city project for Tehran Urban Planning & Research Center

Education

University of Ottawa, M.A in Economics, Major: Data Science, Graduated December 2018 Allameh Tabatabai University, M.A in Economics, Major: Financial Econometrics, Graduated February 2015 University of Tehran (Transfer Student), B.S in Industrial Economics, Graduated June 2013 Workshops and Certificates

Big Data Specialization (Introduction to Big Data, Graph Analytics for Big Data, Big Data Integration and Processing, Big Data Modeling and Management Systems, Machine Learning with Big Data, Big Data-Capstone Project), Coursera (University of California, San Diego). Hadoop Platform and Application Framework, Coursera (University of California, San Diego). SQL for Data Science, Coursera (University of California, Davis). Machine Learning (in progress), Coursera (Stanford University). Deep Learning Specialization (Neural Networks and Deep Learning, Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization, Structuring Machine Learning Projects, Convolutional Neural Networks) (in progress), Coursera (deeplearning.ai). Google Cloud Platform Big Data and Machine Learning Fundamentals (in progress), Coursera (Google Cloud) Applied Data Science with Python Specialization (Introduction to Data Science in Python, Applied Plotting, Charting & Data Representation in Python, Applied Machine Learning in Python, Applied Text Mining in Python) (in progress), Coursera (University of Michigan). Machine Learning and Reinforcement Learning in Finance Specialization (Guided Tour of Machine Learning in Finance, Fundamentals of Machine Learning in Finance, Reinforcement Learning in Finance, Overview of Advanced Methods of Reinforcement Learning in Finance) (in progress), Coursera (New York University Tandon School of Engineering) Algorithmic Trading Using MATLAB, Iran Finance Association Big Data Economics, Khatam University.

Application of Stata in Economics, University of Allameh Tabatabaei Application of Neural Networks in Economics, University of Tehran. Application of Artificial Intelligence in Economics, University of Tehran Volunteer and Other Activities

Data for Good Ottawa, As a Volunteer Data Scientist (2018). Secretary of Student Scientific Chapter (SSC), Department of Economics at University of Tehran (2010-2012). Member of the Executive Board of Economists of Future Association, Department of Economics at University of Tehran (2010-2013). Member of the Executive Board in 5th & 6th Iranian Tax and Fiscal policies conference (2012-2013), Member of the Executive Board in The 1st Economics Conference in Iran (2011)