Rutgers Master Courses: of Statistics Business Quantitative and School Finance Machine (GPA: Learning, 3.65) Derivatives, Financial Timeseries, Object Oriented Newark, Programming NJ (Aug. 2019 (C+– + May. & Python)2021), Stochastic Calculus, Risk Management, Numerical Analysis Arizona M.Sharif Sc., Construction University State University Management of Technology (GPA 3.92), Top 5% class ranking Tehran, Tempe, Iran AZ ((Aug. Aug. 2013 2008 – – Dec. Jun. 2014) 2013) B.EXPERIENCE Sc., Civil Engineering, (GPA: 3.61)
American Express Artificial Intelligence Lab New York, USA (May 2020 – Aug. 2020) Data Science Intern
Tehran • • • Developed learning Worked information. Developed Stock o o Exchange Generated measures Implemented and models with and the finding big analyzed first such and data - a potential phase Self-as number the then tables Decision social funded social finalizing of fraudsters, of Big networks utilizing network stock Tree, new Data and attributes Random trader Hive, Artificial and of inactive uploading Amex new Python, Forest, customers, customers attributes for Intelligence these each SQL, and attributes customer Regression. for with and and Explainability different credit Spark more as and a than bust to modeling new for extract, outs 65M Tehran, table by each that implementing nodes purposes on analyze, community improved Iran Amex and (such 10M Data Jan. and accuracy such different 2016 as communities. warehouse. report fraud as – Aug. centrality customer detection 7%machine 2019) . Iran Tazand • • • Railways Outperformed Gathered causes Optimized Consulting driving - Researcher data decision Group the operational and benchmark making implemented - Contract inefficiencies. by on ranking average Specialist different the by main statistical 6% over problems 3 and years & machine needs by long-of learning only the projects portfolios Tehran, Tehran, methods improving of Iran stocks, to Iran identify (Mar. (project Jul. bonds 2017 2015 main and speed – – derivatives Aug. underlying Jun. by 2017) 2019) 25%. RELEVANT Predicting • Identified loan ACADEMIC defaults new potential from PROJECTS customers lending and club established database database capturing 17 new clients exceeding forecast by 15%.
• Cleaned regression, the support data, did vector exploratory machine, data decision analysis, tree, implemented random forest, different and neural machine networks learning and compared models such the as results logistic to choose the best model regarding the accuracy and AUC score. IMDB review classification
• Used Lasso and Ridge shrinkage methods, Naïve Bayesian Rule, and Vector Space models.
• Recognized the most positive and negative reviews and the most impactful words SKILLS, CERTIFICATES AND AWARDS
• • • Programming Financial Carlo Certificates: with Python Simulation, and (Quantitative DataCamp) Modeling and Machine Data / Analytic skills: Analyst Learning Quantitative Bond with tools: algorithms, Valuation, R Finance Python, (DataCamp) NLP, & Portfolio SQL, Algorithmic Sentiment Tableau, / Data Analysis, Analyst Trading Git, Analysis Bash Black-with in Scripting, Scholes-Python Python Certification Merton (C+DataCamp) +, R, & Spark, Option (/ Udemy) Machine Hive, Pricing, MATLAB / Learning Unix Monte- and Bash Querying scripting tool, Hive (Udemy) (Udemy) / AWS / Bloomberg cloud services Market (EDX) Concepts / Spark and Certification Python for (Bloomberg) Big Data with / Git Pyspark complete: (Udemy) The definitive, / Hadoop PUBLICATION Hossein • • step-CFA Brilliant Vashani, Level1 by-step Talented Jera AND Candidate guide Sullivan, AWARD Student to Git (March and (Udemy) Title Mounir 2021) of 2008 / Tableau El for Asmar. being 2020 "ranked DB A-Z 2020: (Udemy) among Analyzing the top 0.and 1% of forecasting 350,000 participants design-build in Iran market trends" Journal of Construction Engineering and Management 142.6 (2016): 04016008.
• Extracted, means clustering cleaned, Data and Mining reconstructed method data was done of top with 100 R construction Statistical Software companies’ to determine revenues for different different clusters sections whose of the revenues market. K-
• were Valuation Series similar models of construction during using the R and 11-companies Jump year software study and period. prediction of each cluster market share change by implementing Regression and Time-