Hossein Kalkhoran, Data Scientist and Developer in Vancouver, BC, Canada
Hossein Kalkhoran

Data Scientist and Developer in Vancouver, BC, Canada

Member since October 29, 2021
Hossein is a data scientist passionate about finding hidden patterns in large-scale data. He built an automatic vehicle tracking system using state-of-the-art deep learning algorithms for a large, multistory car park and designed an automated testing and submission system for the University of British Columbia. Combined with strong analytical, project management, and problem-solving skills, Hossein helps technology companies set up, build, test, and optimize their machine learning models.
Hossein is now available for hire




Vancouver, BC, Canada



Preferred Environment

Linux, Windows, R Studio, Google Cloud, Amazon Web Services (AWS), AWS, Python

The most amazing...

...project I've developed is an AI-powered chat system capable of classifying, filtering, and moderating text-based and image-based human interactions.


  • Marketing Data Scientist

    2021 - PRESENT
    Fortune 500 Company (Toptal Client)
    • Designed and set up randomized paired geo experiments for estimating incremental return on ad spend (iRoAS).
    • Set up and deployed statistical models on Google Cloud Platform using services such as Google BigQuery.
    • Developed reports, visualizations, and key performance indexes to track the effectiveness of marketing and guide marketing decisions.
    Technologies: Google Cloud Platform (GCP), Statistical Methods, Hypothesis Testing, Python, SQL, Docker
  • Senior Data Scientist

    2021 - 2021
    Toptal Client
    • Developed a risk assessment model associated with users' withdrawal and deposit limits on a crypto exchange platform.
    • Designed the cloud architecture for deployment of the model into production.
    • Created cloud architecture for monitoring the deployed model in the production.
    Technologies: Amazon Web Services (AWS), Python, Jupyter, Cloud Architecture
  • Senior Data Scientist

    2020 - 2021
    System Toose co.
    • Led a team of four data scientists and three software developers.
    • Designed and built an AI-powered chat system capable of classifying, filtering, and moderating text-based human interaction.
    • Designed an AI-powered image recognition system capable of classifying sensitive or inappropriate images.
    • Developed a time-series anomaly detection service that helps customers monitor various metrics. We took advantage of a simple yet strong, deep learning algorithm.
    Technologies: Big Data, Python, Natural Language Processing (NLP), Image Processing, AWS, Machine Learning
  • Data Scientist

    2019 - 2020
    System Toose co.
    • Worked collaboratively in a team of deep learning researchers.
    • Designed and implemented a big data pipeline using Apache Hadoop for analyzing over 350 GB of data.
    • Built an automatic vehicle recognition and tracking system using state-of-the-art deep learning algorithms. The system was capable of detecting and tracking vehicles within a large multistory car park.
    Technologies: Data Science, SQL, Modeling, Predictive Modeling, Data Visualization, Tableau, Big Data
  • Software Developer

    2019 - 2020
    The University of British Columbia
    • Built a REST API back end for a learning analytics website.
    • Designed and implemented an automatic testing and submission system for university students to increase the overall quality of computer science courses.
    • Installed Linux and virtualized environments using Docker and Amazon Web Services.
    Technologies: Back-end, Django, Amazon EC2, Docker
  • Data Analyst

    2018 - 2019
    Mad Llama Studio
    • Developed a data analysis dashboard capable of analyzing a real-time data stream and generating appropriate reports.
    • Designed and developed a data cleaning pipeline specific to our internal processing tasks.
    • Improved one of our client's eCommerce websites using predictive models. We optimized the website's performance metrics as 1) bounce rate: 21% decrease, 2) average session duration: 51% increase, and 3) pages and sessions: 18% increase.
    • Managed to increase the sales volume of a client's online business by 2,000% in a 2-month project.
    Technologies: Pandas, Scikit-learn, PyTorch, Keras, NumPy, Tableau, Google Ads, Facebook Ads, Digital Marketing
  • Data Analyst

    2015 - 2018
    Motamed Cancer Institute
    • Worked in the microbiome and bioinformatic lab as a research assistant.
    • Designed and implemented a new mathematical model to predict the actual. drug release from a specific type of biomaterial.
    • Designed and build data pipelines using Python and R for analyzing and optimizing experimental models.
    Technologies: Python, Data Visualization, Mathematical Modeling, R Studio, Data Analyst


  • Content Moderation System

    A system for moderating text and image-based content on social media. Using natural language processing and image classification techniques, we designed predictive models capable of classifying, filtering, and moderating inappropriate text and image messages. In this project, I was the lead data scientist. My responsibilities included preparing our training dataset, designing the architecture of the models, leading a team of data scientists to train and optimize our models, and testing and evaluating our results.

  • Real-time Anomoly Detector

    A system for unsupervised anomaly detection in multivariate time series data. This project was developed using a convolutional neural network and spectral residual to detect anomalies in time-sensitive data streams.

  • Traffic Monitoring System

    A traffic monitoring system for a large multistory car parking company. I designed a multi-camera vehicle detection and tracking system. It was capable of vehicle re-identification in the client's multi-camera video surveillance system. The objective of the project was to track the activities of drivers on the premises.


  • Languages

    Python, R, SQL
  • Libraries/APIs

    Scikit-learn, PyTorch, NumPy, TensorFlow, Pandas, Keras
  • Tools

    Tableau, R Studio, Jupyter, Amazon SageMaker
  • Paradigms

    Data Science
  • Platforms

    Linux, Amazon Web Services (AWS), Google Cloud Platform (GCP), Windows, Amazon EC2, Docker
  • Other

    Machine Learning, Big Data, Data Analysis, Predictive Modeling, Quantitative Analysis, Natural Language Processing (NLP), Image Processing, Data Visualization, Google Ads, Facebook Ads, Digital Marketing, AWS, Back-end, Convolutional Neural Networks, Modeling, Mathematical Modeling, Data Analyst, Cloud Architecture, Image Segmentation, Statistical Methods, Hypothesis Testing, Recommendation Systems, Classification Algorithms
  • Storage

    Google Cloud
  • Frameworks

    Hadoop, Django, MXNet


  • Master's Degree in Computer Science
    2019 - 2021
    University of British Columbia - Vancouver, Canada
  • Bachelor's Degree in Chemical Engineering
    2012 - 2016
    Iran University of Science and Technology - Tehran, Iran


  • Object Detection with Amazon Sagemaker
  • Image Classification with Amazon Sagemaker
  • Building Recommendation System Using MXNET on AWS Sagemaker
  • Semantic Segmentation with Amazon Sagemaker

To view more profiles

Join Toptal
Share it with others