Dedicated to learning new skills and striving to achieve personal and educational goals.
Currently a MSc student at Polytechnique Montreal(UdeM). Looking for a job as a data scientist or machine learning engineer.

---

Publications

Explaining the behavior of reinforcement learning agents using association rules | Fall 2022

- Deep reinforcement learning algorithms are increasingly used to drive decision-making systems. However, there exists a known tension between the efficiency of a machine learning algorithm and its level of explainability.
Generally speaking, increased efficiency comes with the cost of decisions that are harder to explain. This concern is related to explainable artificial intelligence, which is a hot topic in the research community.
In this paper, we propose to explain the behaviour of a deep reinforcement learning algorithm thanks to standard data mining tools, i.e. association rules. We apply this idea to the design of playing bots, which is ubiquitous in the video game industry.
To do so, we designed three agents trained with a deep Q-learning algorithm for the game Street FighterTurbo II. Each agent has a specific playing style. Our experiments show that association rules can provide interesting insights on the behavior of each agent, and reflect their specific playing style. We believe that this work is a next step towards the explanation of complex models in deep reinforcement learning.

Projects & Experiences

Hockey Goal Prediction | University of Montreal | Fall 2022
- We analyzed hockey data, specifically the NHL stats API, in order to retrieve data for a specific period of time using REST API concepts.
- We have formatted the acquired data in a way that makes it easier to analyze the games.
- We created interactive and representative plots using matplotlib, Seaborn, and ploty to analyze and gain insights into the games and communicate the results more effectively.
- Designed and created downstream features, encoded features not directly usable in ML models, and created a good xG classifier.
- We tracked our experiments using Comet.ml.
- A model service flask app has been built using all the models that were trained and saved into Comet ML and deployed into a Docker container.
- In addition, we built a “Live game client” that retrieved shot events from live (or historical) NHL games, preprocessed them into features compatible with our model, and then queried our model service for expected goal predictions.
- Streamlit has been used to package the live game client into an interactive dashboard.
- You can view the deployed final project here. You can select the model and version, as well as the game ID to get the prediction.

Kaggle Competition | Polytechnique Montreal | 2021
- implemented an efficient algorithm to classify a given set of animal images into 11 classes.
- The report for this project can be found here.

Regularization and Feature Selection in Least Squares Temporal Difference Learning | Polytechnique Montreal | 2021
- Python-based implementation of Least Angle Regression Temporal Difference (LARS-TD) algorithm and Least-Squares Temporal Difference (LSTD).

Machine Learning Researcher Intern | StockholmSyndrome.ai | 2021
- Developed an algorithm to explain the behavior of a deep reinforcement learning algorithm thanks to standard data mining tools, i.e. association rules.
- Designed three agents with specific playing styles trained with a deep Q-learning algorithm for the game Street FighterTurbo II. A designed algorithm provided insight into each agent’s behavior and demonstrated their specific playing style.

Data Developer | Snapp Company | 2019
- Wrote SQL queries to aggregate data.
- Performed business analysis and wrote SQL scripts to analyze data and parse it to Excel.
- Addressed ad-hoc to provide visible data for data analysts and businesses using SSRS and Power Bi.
- Developed ETL pipeline in python.
- Developed Clickhouse Data Pipeline for Real-Time Analysis and Reporting.
- Developed Powerbi/Superset/Grafana Tech Stack for Visualization and advanced reporting and dashboard analysis.
- Worked on Clickhouse, MySql, SQL Server, and BigQuery.

Multi-Agent Deep Reinforcement Learning | Bachelor’s Thesis | 2019
- Addressed pursuit-evasion problem with a multi-agent deep reinforcement learning algorithm.

---

Education

University of Montreal/ Polytechnique Montreal| MSc in Computer Engineering | 2021-2023
Amirkabir University of Technology| BSc in Computer Science | 2015-2019

---