Hello there, I'm Ashot

Work Experience

NLP-egineer at Coelus, Nov 2022 - now:
- Developed graph algorithms and data structures to extract information from text.
- Increased the speed of data preprocessing using multiprocessing.
- Improved code quality through the use of principles of code development (SOLID, KISS, DRY, etc).
- Decreased volume of space consumed by serialized files through the use of an efficient method of storing data.
Data Scientist at scientific research, April 2024 - now:
- Trained neural net for semantic segmentation of liver cells.
- Developed demo version with GUI.

2023-2025, Masters Degree in Applied Mathematics at MIREA - Russian Technological University
2019-2023, Bachelor Degree in Applied Informatics at MIREA - Russian Technological University
- Machine Learning, Deep Learning, Database Management Systems, Statistics, Linear Algebra, Business Process Modeling.

Face Recognition & Voice Greeting system - developed a voice greeting system based on a recognized face for the institute in the team
Stereo Depth Estimation - developed neural net for fast and memory efficient stereo depth estimation
PICA - web site for style transfer, more detailes in code
Tiling problem solution - python application for tiling problem solution, details are in repo.
HabrParser - I parsed habr and make some analysis: word clouds, engagement rate calculation and correlation analysis.

Development of algorithms for generating voice deepfakes, 3rd place:
- Implemented algorithm for cleaning the speaker's audio recordings from noise.
- Removed silence from audio files.
- Made GUI demo application.
WoodHack: classification of wood images into classes grade 1, grade 3 and firewood, code