Dinko Vasilev, Machine Learning Developer in Kiev, Ukraine
Dinko Vasilev

Machine Learning Developer in Kiev, Ukraine

Member since March 3, 2018
Dinko is a data scientist with a strong background in machine learning algorithms, mathematics, statistics, and programming. For over six years he has been searching for the signal in piles of data, refining it, understanding it, and putting it to use. The bulk of his work for the past few years has consisted of machine learning projects but also some statistics.
Dinko is now available for hire

Portfolio

  • Freelance Work
    Python, R, AWS, Docker, SQL, NoSQL
  • Anagog
    Python, R, SQL, Amazon Web Services (AWS), Docker
  • SoftBistro
    Python, R, SQL, Amazon Web Services (AWS), Docker

Experience

  • Python, 8 years
  • Statistics, 6 years
  • R, 6 years
  • Machine Learning, 6 years
  • Data Science, 6 years
  • Natural Language Processing (NLP), 4 years
  • Computer Vision, 4 years
  • AWS, 4 years

Location

Kiev, Ukraine

Availability

Part-time

Preferred Environment

Python, R

The most amazing...

...product I've worked on is a model that predicts the outcomes of UK horse races.

Employment

  • Freelance Data Scientist

    2018 - PRESENT
    Freelance Work
    • Developed a churn prediction model for a large US-based software company. The model predicts the probability that a customer would churn three-to-six months ahead in time.
    • Modeled the result of a clinical trial, specifically patients' reactions to the medication.
    • Built a model that predicts equipment failure for a large manufacturer.
    • Developed a trading strategy for a Hong Kong-based cryptocurrency hedge fund.
    • Used NLP to build a product that recommends filmmakers. It processed the titles, descriptions, and metadata of millions of video posts.
    • Used object detection and image classification on images of restaurant meals to detect, classify, and estimate the number of ingredients.
    Technologies: Python, R, AWS, Docker, SQL, NoSQL
  • Data Scientist

    2018 - 2019
    Anagog
    • Used NLP to create an app that detects addresses in webpages.
    • Built a tool that automatically detects and presents insights/interactions of interest in numerical data..
    • Developed models that detect and classify smartphone users' activities based on sensor data.
    • Modeled the time interval between waking up and leaving for work based on smartphone sensor data.
    • Used NLP to detect and classify points of interest using Common Crawl data.
    Technologies: Python, R, SQL, Amazon Web Services (AWS), Docker
  • Data Scientist

    2017 - 2018
    SoftBistro
    • Predicted customer orders for a large US-based wholesaler.
    • Developed an algorithm that determines the optimal product mix for a customer given constraints.
    • Built an image recognition pipeline that detected, classified and counted products(snacks, drinks, and the like) on shelves and fridge shelves.
    Technologies: Python, R, SQL, Amazon Web Services (AWS), Docker
  • Data Scientist

    2016 - 2017
    Thoroughbet
    • Improved the profitability of a model that predicts the outcome of UK horse races.
    • Generated new features.
    • Optimized model hyperparameters.
    • Devised new strategy for dealing with missing data.
    • Theoretically modeled the tendency of a model to overfit and compared it to the empirical results.
    Technologies: Python
  • Data Scientist

    2013 - 2016
    Startup
    • Developed a trading strategy backtesting engine.
    • Built short-term forecasting models.
    • Developed rules that translated forecasts into signals.
    • Created an online algorithm for capital allocation.
    • Built a portfolio risk model.
    Technologies: R, Python, Interactive Brokers, SQL, C++

Experience

  • Trading Strategy Development (Development)

    I worked on the trading strategy development for a trading company and more recently cryptocurrency hedge funds.

  • NLP (Development)

    I used NLP on a variety of projects, e.g., processing the textual data of millions of video posts and Common Crawl data.

  • Anomaly Detection, Insight Generation, and Forecasting (via Toptal) (Development)

    I worked on anomaly detection, insight generation, and forecasting on credit card sales data.

  • Computer Vision (Development)

    I developed an image recognition pipeline that detects, classifies and counts products (snacks, drinks, and similar) on shelves and fridge shelves.

  • Computer Vision (via Toptal) (Development)

    I used object detection and image classification on images of restaurant meals to detect, classify and estimate the number of ingredients.

Skills

  • Languages

    Python, R, JavaScript
  • Libraries/APIs

    Keras, Sklearn, XGBoost, NumPy, Tidyverse, Ggplot2, Matplotlib, TensorFlow
  • Tools

    Microsoft Word
  • Paradigms

    Data Science, Quantitative Research
  • Platforms

    AWS EC2, Docker, Linux, Amazon Web Services (AWS)
  • Storage

    AWS S3
  • Other

    Machine Learning, Generalized Linear Model (GLM), Statistics, Econometrics, Supervisory, Natural Language Processing (NLP), Computer Vision, Deep Learning, Neural Networks, Data Visualization, Artificial Intelligence (AI), Topic Modeling, Keyword Analysis, AWS
  • Frameworks

    Flask

Education

  • Bachelor's degree in Physics
    2011 - 2016
    University of Oslo - Oslo, Norway
  • Bachelor's degree in Statistics
    2011 - 2016
    University of Oslo - Oslo, Norway
  • Master's degree in Finance
    2005 - 2006
    Cass Business School - London, UK

To view more profiles

Join Toptal
I really like this profile
Share it with others