Mehdi is available for hire
Hire MehdiMehdi Ghasemi
Verified Expert in Engineering
Mathematics Developer
Location
Edmonton, AB, Canada
Toptal Member Since
July 4, 2019
Mehdi is a mathematician and scientific programmer specializing in applied math and abstract concepts. His scientific background is rooted in optimization, geometry, and computations. Mehdi has been working as a data scientist for years, building statistical and machine-learning models.
Portfolio
Edmonton Police Service
Bayesian Inference & Modeling, Data Visualization, Jupyter, Python, SQL...
University of Saskatchewan
Statistics, Mathematics, Python, Mentorship & Coaching, Applied Research
Government
SQL, Python, Pandas, Keras, Time Series Analysis, Statistics, Scheduling, Flask...
Experience
Availability
Part-time
Preferred Environment
Jupyter, Git, PyCharm, Python, Pandas, Linux, Web Programming
The most amazing...
...project I've worked on is a global optimization method based on a series of SDPs that in theory is capable of solving any given optimization problem.
Work Experience
Senior Scientist | Mathematician
2020 - PRESENT
Edmonton Police Service
- Analyzed and extracted information of historical documents so that the unstructured data becomes semi-structured and digestible by AI algorithms. This task involved image processing, NLP, LLM, and mathematical optimization.
- Made adjustments to police communication branch staffing to meet and improve upon national minimum standards.
- Forecasted the impact of COVID-19 on the police force and its corresponding risk analysis.
- Analyzed the transition of heavy drug users over time. Also, studied the network of drug users and distributors around the city and effective methods for concentrated disruption of drug-related offenses.
Technologies: Bayesian Inference & Modeling, Data Visualization, Jupyter, Python, SQL, Scheduling, Stochastic Modeling, Machine Learning, Trend Analysis, Forecasting, Mathematical Modeling, Natural Language Processing (NLP), Image Processing, Optimization, Computational Geometry, Pandas, Keras, OpenCV, Mathematics, Time Series Analysis, Statistics, GPT, Web Programming, Flask, Model View Controller (MVC), Visualization, Databricks, Large Language Models (LLMs), OpenAI, Artificial Intelligence (AI), Data Analysis, Data Analytics, Data Scientist, Data Modeling, Data Science, Logistic Regression, Statistical Modeling, Consulting, Technical Leadership, Time Series, Data Manipulation, Generative Pre-trained Transformers (GPT), Modeling, Generative AI, Applied Research, Fine-tuning, Conda
Adjunct Professor
2018 - PRESENT
University of Saskatchewan
- Collaborated with other researchers in scientific projects that involve mathematical modeling.
- Co-supervised graduate students in areas related to optimization and computer science.
- Provided scientific consultation to identify, solve, and implement solutions to industrial problems.
Technologies: Statistics, Mathematics, Python, Mentorship & Coaching, Applied Research
Data Scientist
2017 - 2020
Government
- Designed and implemented pipelines to extract specialized datasets out of the administrative database.
- Used government data and analyzed some of the existing practices to find bottlenecks and optimized procedures.
- Employed machine learning to improve decisions made based on standard assessments.
- Evaluated the feasibility of new policies to achieve certain goals by employing time-series analysis and forecasting.
- Made a local and provincial evaluation of initiatives in correction, justice, and child welfare.
Technologies: SQL, Python, Pandas, Keras, Time Series Analysis, Statistics, Scheduling, Flask, Model View Controller (MVC), Visualization, Machine Learning, Artificial Intelligence (AI), Data Analysis, Data Analytics, Data Scientist, Data Modeling, Data Science, Logistic Regression, Statistical Modeling, Time Series, Data Manipulation, Modeling, Applied Research, Fine-tuning, Conda
Consultant
2017 - 2018
The Centre for Forensic Behavioural Science and Justice Studies
- Analyzed the risk assessment tool LSI (Level of Service Inventory).
- Applied machine learning to LSI data to obtain personalized preventive interventions for offenders.
- Organized and managed workshops to expand the professional and scientific network of the Centre.
Technologies: SQL, SPSS, Python, Pandas, Data Analytics, Data Scientist, Data Modeling, Data Science, Logistic Regression, Statistical Modeling, Consulting, Mentorship & Coaching, Data Manipulation, Modeling
MITACS Postdoctoral Fellow
2015 - 2017
University of Saskatchewan
- Organized the Saskatoon Police Predictive Analytics Laboratory.
- Built a mathematical simulation of the missing children phenomenon to identify its deriving factors among youth.
- Researched the optimization and moment problem in the applied context.
Technologies: Python, Optimization, Mathematical Modeling, Pandas, Web Programming, Flask, Visualization, Machine Learning, Data Analytics, Data Science, Logistic Regression, Statistical Modeling
Postdoctoral Research Fellow
2013 - 2014
Nanyang Technological University
- Implemented semidefinite programming UI for SAGE and Python.
- Research advanced polynomial and convex optimization.
- Developed the topological moment problem using functional analysis and real algebraic geometry.
Technologies: Sage, Python, Mathematical Modeling
Experience
SKSurrogate
https://github.com/mghasemi/sksurrogateSKSurrogate is a suite of tools that implements surrogate optimization for expensive functions based on sci-kit-learn. The primary purpose of SKSurrogate is to facilitate hyperparameter optimization for machine learning models and optimized pipeline design (AutoML).
The version of the surrogate optimization implemented here heavily relies on regressors. A custom regressor based on Hilbert Space techniques is implemented, but all sci-kit-learn regressors are accepted for optimization.
Finding an optimized pipeline based on a given list of transformers and estimators is time-consuming. Evolutionary optimization has been implemented to reduce its time instead of global optimality.
The version of the surrogate optimization implemented here heavily relies on regressors. A custom regressor based on Hilbert Space techniques is implemented, but all sci-kit-learn regressors are accepted for optimization.
Finding an optimized pipeline based on a given list of transformers and estimators is time-consuming. Evolutionary optimization has been implemented to reduce its time instead of global optimality.
Irene Project
https://github.com/mghasemi/IreneIrene is a Python package that aims to be a toolkit for global optimization problems that can be realized algebraically. It generalizes Lasserre's Relaxation method to handle theoretically any optimization problem with a bounded feasibility set. The method is based on solutions of generalized truncated moment problems over a commutative real algebra.
InventoryOptim
https://github.com/mghasemi/InventoryOptimGiven inventory data of multiple interacting commodities from stock with limited but variable capacity, provide insight on:
1. Estimating future required capacity for each item based on a particular terminal segment of data
2. Future cost estimation for each item
3. How the trends of individual items would change, assuming a trend change at given times in the future for some items
4. Given a budget limit, how should trends change to ensure a non-negative residual
1. Estimating future required capacity for each item based on a particular terminal segment of data
2. Future cost estimation for each item
3. How the trends of individual items would change, assuming a trend change at given times in the future for some items
4. Given a budget limit, how should trends change to ensure a non-negative residual
pyProximation
https://github.com/mghasemi/pyProximationThis package was originally written to solve systems of integro-differential equations via the collocation method in an arbitrary number of variables. The current implementation of the method is based on a finite-dimensional orthogonal system of functions. Therefore, additional modules were required to achieve this goal.
Nonlinear Regression
https://github.com/mghasemi/nonlinear-regressionA small suite of tools to perform nonlinear regression, sci-kit-learn style. It uses linear regression and data transformation to perform unweighted nonlinear regression and implements a version of function spaces as Hilbert spaces to do weighted nonlinear regression.
Skills
Languages
Python, SQL, PHP
Libraries/APIs
Scikit-learn, Keras, Pandas, Sage, OpenCV
Tools
LaTeX, PyCharm, Jupyter, SPSS, Git
Paradigms
Data Science, Model View Controller (MVC), Agile Project Management
Other
Mathematics, Mathematical Modeling, Optimization, Machine Learning, Time Series Analysis, Data Analysis, Data Analytics, Data Scientist, Data Modeling, Logistic Regression, Statistical Modeling, Data Manipulation, Modeling, Scheduling, Trend Analysis, Forecasting, Web Programming, Visualization, Data Visualization, Bayesian Inference & Modeling, Natural Language Processing (NLP), Computational Geometry, Artificial Intelligence (AI), Consulting, Mentorship & Coaching, Time Series, Applied Research, Fine-tuning, Image Generation, Text to Image, Conda, Statistics, Stochastic Modeling, Image Processing, GPT, Global Project Management, Deep Learning, Large Language Models (LLMs), OpenAI, Leadership, Technical Leadership, Generative Pre-trained Transformers (GPT), Generative AI
Frameworks
Flask
Platforms
Linux, Databricks
Storage
MySQL, SQLite
Education
2009 - 2012
Ph.D. in Mathematics
University of Saskatchewan - Saskatoon, Canada
2002 - 2004
Master of Science Degree in Mathematical Logic
Tarbiat Modares University - Tehran, Iran
1998 - 2002
Bachelor of Science Degree in Mathematics
Amirkabir University of Technology - Tehran, Iran
Certifications
SEPTEMBER 2023 - PRESENT
Stable Diffusion AI Art
Udemy
DECEMBER 2022 - PRESENT
Project Management Fundamentals
INEXERTUS
Collaboration That Works
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
1
Share your needs
Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2
Choose your talent
Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3
Start your risk-free talent trial
Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.
Top talent is in high demand.
Start hiring