Alexander Khrulev

Experience

• 

  Telecontact

  Jun 2019 - Aug 2019

  Software Developer

  • Wrote automated tests in Java using Selenium Framework and identified 4 critical bugs in software• Achieved full coverage of multiple websites with automated tests• Wrote queries to the MariaDB to automate test data preprocessing• Created weekly reports for the supervisor• Ensured that the software could be deployed on other machines via Maven• Debugged the legacy codeRelevant Technologies:• Java• Selenium Framework• Maven• SQL• git• Jira• Linux(Ubuntu)• Eclipse Show less

• 

  Ryerson University | Laboratory of Innovations in Transportation

  May 2020 - May 2021

  Research Assistant

  • Created a data pipeline for OpenStreetMap data in Python• Transformed and extracted relevant information about road networks using pandas and numpy• Visualized the data and made reports with matplotlib and seaborn* Wrote a code in C++ to use the data to model the Toronto road network on a cluster containing 24 Raspberi Pi* Updated the documentation and debugged legacy codehttps://litrans.ca/

• 

  GEOTAB

  May 2021 - Aug 2022

  Data Analyst Intern

  • Produced various statistical analyses and reports to stakeholders(for ex. the impact of Vancouver mudslides in 2021, SH-130 highway analysis)• Used SQL with Google BigQuery to extract insights from large volumes of data(>1 TB on average)• Wherever feasible(due to the size of data), used Pandas, Matplotlib, and Seaborn to process the data and produce visualizations in Jupyter Notebook/Lab• Produced statistical reports to stakeholders saving a few million of dollars in potential revenue losses(ex. the impact of Vancouver mudslides in 2021, SH-130 highway analysis)• Added new functionality and improved internal project tool using Bash and Python• Created a tool to automatically identify discrepancies between two sources of data that helped to correct the uploader issues on the server side• Created and managed multiple data pipelines in Apache Airflow to process TBs of data• Conducted multiple tutorials and lunch & learn sessions to a large audience (for ex. asynchronous & parallel programming in Python, caveats of Python, basics of Bash)Relevant Technologies:• Python(asyncio, Pandas, Matplolib, Seaborn, Scikit-learn, StatsModels)* SQL* Google Cloud Platform, Google BigQuery* Apache Airflow* Bash Show less

• 

  Cybersecurity Research Lab (CRL)

  Sept 2022 - Apr 2023

  Research Assistant

  • Assisting with cybersecurity corpus text analytics research (primary from United States Patent and Trademark Office)• Wrote wrappers to conveniently use USPTO APIs• Created a script to automatically download new patent data from USPTO asynchronously (with asyncio, aiohttp, aioflies) that significantly reduced the preprocessing time• Made optimizations to minimize RAM consumption of certain scripts(currently tops 20 MB vs a few GB before)• Developed and trained various ML models to analyze unstructured textual data• Achieved 100% documentation coverage• Dealt with legacy Java codeRelevant Technologies:• Python (doctest, asyncio, aiohttp, aiofiles, pandas, jupyter, sklearn)• Bash• Rust• Git• Java Show less

• 

  Ryerson University

  Oct 2022 - Jan 2023

  MTH380 Probability and Statistics I Crash Course Tutor
• 

  Cybersecurity Research Lab (CRL)

  Jul 2023 - Aug 2023

  Research Assistant

  • Researching practical applications of LLMs (Large Language Models) in Cybersecurity Field• Fine-tuning BERT/LLAMA models(via HuggingFace)• Conducted Topic Modelling in Python(BERTopic, Pytorch, Gensim)Relevant Technologies:• Python (Doctest, Pytest, Jupyter, Pandas, Numpy, Gensim, Spacy, Pytorch, Transformers)• Bash• Rust• Git

• 

  Financial Mathematics Research Lab

  Sept 2023 - now

  Research assistant

  • Currently working on Improving Model Fairness with the help of GANs(Generative Adversarial Networks)• Conducting Hyper parameter tuning with Bayesian-optimization(both manually and via Ray)• Following CI/CD practices with Github Actions and Docker• Testing and deploying models on a cluster(Great Expectations, DVC, Weights&Biases)

• 

  Leap Tools

  Mar 2024 - Mar 2024

  Data Scientist

  - Compared an Experimental Conversion Rate (CVR) formula with the current one by analyzing data from 400+ businesses over multiple years, leading to a more accurate performance metric (AWS, Amazon Redshift, Python).- Analyzed CVR calculations across different timeframes (1 to 6 months) and identified the optimal duration, resulting in a 20% improvement in metric accuracy.- Conducted research into the reasons behind high/low CVR, identifying over 10 high-risk customers, which led to targeted retention strategies improving customer retention by 15%.- Proposed and calculated four additional KPIs, enhancing insight extraction and improving strategic alignment by 25%. Show less

• 

  GEOTAB

  Apr 2024 - now

  Software Developer, MLOps

  • Automated the deployment of ML models using Gitlab CI/CD, Docker, and Kubernetes, streamlining the CI/CD process.• Implemented logging, monitoring, and reproducibility workflows with MLFlow and Apache Superset, improving model transparency and debugging.• Managed data pipelines in Airflow, adding multiregional support to increase reliability and reduce latency.• Created and standardized model training/deployment templates, which are now used by over 10 teams, resulting in consistent and efficient workflows. Show less