I am a third-year student at ITMO University, majoring in Robotics and Artificial Intelligence, and I work as an ML engineer. My tech stack includes Docker and Python with libraries like PyTorch, scikit-learn, OpenCV, transformers, FastAPI, unittest. Additionally, I utilize Roboflow and Label Studio for data annotation in specific machine learning tasks. My dream is to become a Kaggle GrandMaster.
Dmitry Zolotarev
Russia Saint-Petersburg
+7 (927) 048 22 12
dmitryellison@gmail.com
My team and I worked on a smart mirror project. To solve the problem, we have trained segmentation models((UNet, YOLOv8) and generatively competitive networks (GAN, GAN-GP).
Together with my classmates, we presented a paper on the project Generative Models in the Fashion Business at the Young Scientists Congress.
I participated in a summer school as part of the team Computer Vision for Oil and Gas Companies. ML backend: data annotation, preprocessing, and augmentation. Model selection and training in PyTorch. The final result of our team's work was an MVP (Minimum Viable Product).
I am involved in the development of internal solutions for the TATNEFT group of companies. In this position, I'm responsible for selecting, training, and deploying models, connecting the backend with the frontend to simplify and improve user experiences.
At the moment, I am studying in the third year of the Robotics and Artificial Intelligence faculty. I have an average GPA of 4.39 out of 5 for the first year and 4.75 for the second one.
I successfully completed the advanced course with a score of 91%. I worked with segmentation, detection, autoencoders, and generative adversarial models. In addition, I became acquainted with classical machine learning models and learned about transfer learning.
This course is dedicated to the use of neural network models for processing natural language and audio signals.
I have implemented a Zero-Shot Classification model from Hugging Face for the task of categorizing key skills on resumes extracted from the hh.ru website.
The application of OCR models for the purpose of recognizing text in photos.
Face recognition using a ResNet encoder and a Random Forest model on a small dataset for a Kaggle competition.
I've successfully applied Neural Style Transfer using pure PyTorch.
I've implemented a face generation system using a DCGAN (Deep Convolutional Generative Adversarial Network) with additional enhancements for improved image quality and realism.
I've developed a digit recognizer using Lasso and Ridge regularization techniques for improved accuracy and robustness."
I've completed a portion of the AGNI workshop, specifically focusing on glade segmentation.