Daniil Orel

Experience

• 

  Университет ИТМО

  Dec 2019 - Feb 2020

  AI Research Assistant

  Worked towards investigation of research preferences of the University profs.Used data scrapping techniques (bs4, selenium) to parse more than 5K research papers;Developed an NLP classifier to recognise the research field by the paper abstract

• 

  AEROSPACE-AGRO

  Oct 2020 - May 2021

  ML engineer

  Conducted an early data analysis of 10K+ satellite images to detect corrupted data. It helped to build a dataset for further use. Represented findings to the team leader and lead data analyst to form teams for data correction.Built a baseline model for Segmentation of Satellite Images, using DeepLab3. IoU score -0.66.This project got intoTOP-30 best projects utilizing AI among 1000 participants in ”Archipelag” accelerator.Developed a fully-convolutional neural network for segmentation of hogweed on satellite images (IoU score: 0.58) Show less

• 

  ISSAI

  May 2021 - Feb 2022

  AI Researcher

  Studied Language Identification techniques and developed multiple models for language identification. Applied domain adaptation and augmentation techniques for audio data.Key characteristics of the models:- In-domain data: F-score = 0.93 - Out-of-domain data: F-score: 0.41 (another model)

• 

  Yandex

  Dec 2021 - Apr 2023

  Data Scientist
• 

  Google Summer of Code

  Jun 2022 - Sept 2022

  CV engineer (VR)

  I have accomplished the Google Summer of Code program, contributing to LibreHealth's project.My task was to develop a set of CV models that can operate in VR/AR environment. As a result, I come up with an augmentation which can simulate VR distortion. Additionally I have:* Developed Classification/Detection and Segmentation models that can operate in VR.* Optimised the models to run on CPU* Created a class for easier video-inference of modelsMore details are here: https://daniilor.blogspot.com/2022/09/summing-up-my-gsoc-experience.html Show less

• 

  MBZUAI (Mohamed bin Zayed University of Artificial Intelligence)

  May 2023 - Jun 2023

  Research Assistant (NLP)

  Developed a system for News-source bias and factuality evaluation and deployed it.* Researched SOTA approaches to bias and factuality evaluation* Implemented Sentence-Transformers for ordinary regression problem* Designed 2-level model for factuality and bias evaluation, which gives TOP-3 result on CLEF 2023 CheckThat! lab's task* Deployed the whole system