Tornike Onoprishvili, Developer in Tbilisi, Georgia
Tornike is available for hire
Hire Tornike

Tornike Onoprishvili

Verified Expert  in Engineering

Data Scientist and Developer

Tbilisi, Georgia

Toptal member since October 1, 2021

Expertise
PythonMachine LearningTensorFlowMathematicsComputer VisionNLPSQLSoftware DevelopmentData ScienceBig Data Architecture
Bio

Tornike is a data scientist with a solid academic and mathematical background. He specializes in Python TensorFlow 2.0, currently learning reinforcement learning, and is planning to contribute to this field in the near future. Tornike is also regularly busy mentoring aspiring data scientists on their path to greatness.

Portfolio

Freelance
Python, Jupyter Notebook, TensorFlow, Scikit-learn
ACT - Analysis and Consulting Team
Python, Pandas, SQL, Scikit-learn, XGBoost
Orient Logic
Python, KNIME, Matplotlib, Presentations, Microsoft PowerPoint, Microsoft Excel...

Experience

  • Python - 4 years
  • Machine Learning - 4 years
  • Data Science - 4 years
  • Keras - 3 years
  • Pandas - 3 years
  • Jupyter Notebook - 3 years
  • TensorFlow - 3 years
  • Google Cloud ML - 2 years

Availability

Part-time

Preferred Environment

Jupyter Notebook, Python, TensorFlow, Pandas, Scikit-learn, Google Cloud ML, SQL, Machine Learning, Data Science

The most amazing...

...achievement was winning second place in Tox21 challenge by intelligent feature engineering techniques.

Work Experience

Data Scientist

2021 - PRESENT
Freelance
  • Designed a deep learning architecture for fully-automatic bio-assay toxicity prediction based on Mol2vec (Jaeger et al., 2018) preprocessing and achieved the best single-model test performance on the Tox21 sub-dataset.
  • Created a trading analytics pipeline for automatic gathering, sorting, and combining various trading algorithms with Python.
  • Implemented and deployed an artistic style transfer network for the stylization of everyday images.
  • Designed and implemented a Python scheduled data delivery script for automatic data gathering and snapshotting.
  • Created an AI-based VR application for predicting user stress levels based on the pupil diameter for Pico Neo 2.
Technologies: Python, Jupyter Notebook, TensorFlow, Scikit-learn

Mid-senior Data Scientist

2020 - 2021
ACT - Analysis and Consulting Team
  • Created a collaborative filtering solution for automatic recommendations based on customer transaction data of a local pharmaceutical company.
  • Developed a data pipeline for cleaning, aggregation, and automatic reporting for a hospital transactional database.
  • Oversaw communications with the data engineering team and delivered reports based on the data insights.
Technologies: Python, Pandas, SQL, Scikit-learn, XGBoost

Junior Data Scientist

2018 - 2020
Orient Logic
  • Developed a transaction fraud detection system for a local bank with a modified IsoForest-based anomaly detection algorithm, using Scikit-learn.
  • Built an experimental Georgian-native banking assistant chatbot using IBM Watson and Flask. Flask server used NLP to normalize Georgian input using a Google API prior to sending requests to IBM Watson.
  • Worked on and presented promising AI technologies and libraries to the leading managers in order to boost Orient Logic's business analytics capabilities.
Technologies: Python, KNIME, Matplotlib, Presentations, Microsoft PowerPoint, Microsoft Excel, TensorFlow, Scikit-learn

Attentional Style Transfer

Implemented a style transfer paper (Dae Park and Kwang Hee Lee, 2019) using PyTorch, TensorFlow, and TensorFlow.js into a fully serverless web application that can be viewed as a GitHub Pages app. This challenging transition between PyTorch, ONNX, and TensorFlow.js gave me a unique understanding of inter-framework relationships and ML model deployment considerations.

Tox21 Challenge

https://tripod.nih.gov/tox21/challenge/
The Tox21 challenges data scientists around the world to predict chemical interactions with biological pathways. I developed a statistical model—ensemble (RF, GBM, and DT) and tuned the model via Bayesian optimization to address a specific sub-task in this challenge. With 50+ teams competing to build the best predictor, my model was second due to my hand-crafted innovative feature engineering techniques.

Automatic Trading Portfolio Construtor

A Python-based app for gathering the unique trading algorithms in a group for automatic portfolio creation. For over a month, I collaborated with a non-technical client to implement, experiment with, and release an end-to-end Python Pandas pipeline to process large streams of trading data into a diverse trading portfolio.
2015 - 2019

Bachelor's Degree in Mathematics and Computer Science

Free University of Tbilisi - Tbilisi, Georgia

FEBRUARY 2021 - FEBRUARY 2023

TOEFL

Educational Testing Service

Libraries/APIs

TensorFlow, Keras, Pandas, Scikit-learn, XGBoost, Matplotlib

Tools

Microsoft PowerPoint, Microsoft Excel, Open Neural Network Exchange (ONNX)

Languages

Python, SQL

Platforms

Jupyter Notebook, KNIME

Storage

Google Cloud

Other

English, Data Science, Machine Learning, Mathematics, Google Cloud ML, Computer Vision, Natural Language Processing (NLP), Software Development, Statistics, Statistical Analysis, Convolutional Neural Networks (CNNs), Cloud Computing, Big Data, Forex Trading, Stock Market, Presentations, Machine Learning Operations (MLOps), Generative Pre-trained Transformers (GPT)

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring