Data Scientist2020 - 2021ATH Digital LLC
Technologies: Docker, Plotly, PostgreSQL, AWS S3, AWS Lambda, Jupyter Notebook, Pandas, AdWords API, Facebook API, Cron, Python, AWS Kinesis, AWS EC2, Docker Compose, Jupyter, Google Analytics API, Apache Airflow, Big Data, AWS
- Created data ingestion scripts for pulling data from ad platforms like Adwords and Facebook Ads.
- Developed automatic uploading of the CSV and Excel files data into the database based on the AWS services.
- Set up the marketing streaming cloud infrastructure of the data processing pipeline.
- Designed a database model based on the data science team requirements.
- Created a model for forecasting and visualizing the balance burn rate metric.
Senior Data Scientist2019 - 2020Zelos.AI
Technologies: AWS EMR, PySpark, Jupyter, Amazon Web Services (AWS), Statistics, Data Science, AWS DynamoDB, AWS Lambda, AWS EC2, AWS S3, lxml, Data Modeling, Database Modeling, Code Architecture, Markov Model, Markov Chain Monte Carlo (MCMC) Algorithms, Batch, Scrapy, DB, Data Scraping, Selenium, Data Engineering, Machine Learning, Natural Language Processing (NLP), ETL, Docker, AWS, Python, Apache Airflow, Pandas, Big Data
- Processed and analyzed over 100 million athletic performance data with PySpark running on AWS EMR.
- Designed a data model based on the companies business requirements.
- Made a batch data processing pipeline orchestrated by Airflow.
- Created a data scraping tool for parsing dynamic and static web pages using Scrapy, Selenium, lxml.
- Developed athletics competitions simulations based on the Monte Carlo approach.
Data Scientist2018 - 2019Windsor.AI
Technologies: Jupyter, DB, Marketing, Google Analytics, PostgreSQL, SQL, Statistics, R, Pandas, Python, Docker, Facebook API, AdWords API, Big Data, AWS
- Optimized existing SQL queries, making them less complex and having higher performance.
- Used SQL for gaining insights, detecting anomalies and problems in the collected data.
- Created a workflow for the data migration between different database management systems.
- Developed scripts for ingesting data from different online advertising platforms.
- Designed new database tables according to the analytics team requirements.
Data Scientist2018 - 2019Frontier Data Corporation
Technologies: Jupyter, DB, Time Series Analysis, R, Natural Language Processing (NLP), Big Data, Python, Pandas, Docker, PostgreSQL, AWS
- Developed models for trend detection in the Twitter stream.
- Developed AI-based application's architecture.
- Integrated in-house ML models with cloud services as IBM BlueMix and Google Cloud NLP.
- Worked with big datasets using Google BigQuery.
- Created customized modules for new ML models evaluation.
- Trained machine learning models for text classification.
- Created tests for existing applications.
Data Scientist2016 - 2018Pulsar AI
Technologies: Jupyter, DB, MongoDB, Git, Docker, NumPy, Pandas, SpaCy, fastText, Keras, NLTK, Gensim, Scikit-learn, Python, PostgreSQL, AWS, AWS Lambda
- Developed a chatbot framework for the Georgian language applying machine learning and natural language processing (NLP) techniques.
- Trained and deployed a machine learning model for an automated grouping of the news and articles from Georgian media websites.
- Designed a tool for sentiment classification on texts from social networks.
- Analyzed a large amount of user conversations data applying NLP, statistics and presented precise results.
- Worked with time series for analyzing and predicting cryptocurrency prices.
- Managed a team of linguists who worked on the data collection and labeling.
Software Developer Internship2016 - 2016Virtuace Inc.
Technologies: XML, Apache Tomcat, Java, Git, Linux, Docker
- Fixed bugs.
- Expanded functionality of the existing application.
- Tested new modules.
Full Stack Software engineer2014 - 2016Georgian Technical University
- Developed the front-end for managing and working with linguistic corpora.
- Created web services for operating with linguistic corpus data.
- Organized database structure for storing and manipulating the linguistic corpora.
- Analyzed documents using NLP tools and presented results in a clear manner.