English:
Advanced
Viacheslav K.
Vetted by YouTeam
Switzerland
UTC +01:00
Europe/Zurich
English:
Advanced
Machine Learning Engineer
Skilled Machine Learning Engineer and Data Scientist with experience in computer vision, clustering analysis, object detection, and tabular data classification. Also well acquainted with classification, decision trees, data pre-processing, cleaning, as well as neural networks. Have solid experience in AWS services related to ML lifecycles projects, namely inference and training with Sagemaker, data processing via EMR and Glue, and managing projects via ECR and Artifacts.
Want to hire this engineer?
Check if Viacheslav is availableExpertise
Years of commercial development experience
4 years of experience
Core technologies
Other technologies
Project Highlights
Real Estate Market Analytics
Jul `22 - Aug `24
2 years
DataFlik
Predicting probabilities of sell and discounts for properties in the USA over different states.
Responsibilities & achievements
• Led a team of ML Engineers and Data Engineers to collaborate on various initiatives. • Constructed datasets for model retraining, working with over 8TB in a delta lake. • Built and maintained scalable, decoupled, and reliable project infrastructure. • Developed and trained over 2,000 CatBoost models. • Deployed multiple production models with shadow testing for performance evaluation. • Analyzed model performance and extracted feature insights. • Built dashboards to visually represent results and insights.
Image to Text
Mathleaks
Inputs were images and output should be the formulas as a sequence of latex tokens - Image to text generation.
Responsibilities & achievements
• Developed preprocessing and normalization pipelines for LaTeX tokens. •Built a custom Encoder-Decoder model for predictions. •Tested open-source training solutions to assess their effectiveness. •Conducted prediction analysis and performed iterative retraining to improve model accuracy.
Lead Generator
Software Development Company
Having a proposal on Upwork, the task was to generate the relevant lead message to a customer. The message consisted of fields for filling by the text networks and templates given from the DB.
Responsibilities & achievements
• Created templates for lead messages. • Scraped a wide range of data to fulfill templates. • Provided an API in Flask for convenient use by third-party vendors.
Matching System
Jobwallet
Parse and match set of CVs and Job Descriptions using rule-based keyword extraction and machine learning text classification.
Responsibilities & achievements
• Tested various types of classification networks. • Developed a keyword-based rule engine for calculating scores. • Gathered keyword mapping functions across the project. • Created a multi-parser for handling PDF, DOC, and TXT documents. • Implemented assumption systems based on input information.
Person Re-Identification
IT Company
Detecting and tracking people on the video in real-time, identifying them.
Responsibilities & achievements
• Built a pipeline using DeepSort and YOLOv5 for object detection and labeling. • Developed an interactive demo using Flask. • Identified human behavior by tracking movement in stores and on streets.
Pulmonary Fibrosis Detection
Medical Company
Having an unstructured list of patients’ visits info and 3d CT scans of their lungs, the goal was to predict the lungs’ capacity by a certain date and give the probability of the prediction.
Responsibilities & achievements
• Established a 3D data preprocessing pipeline. • Tuned EfficientNets to push metrics for lung capacity prediction. • Conducted post-prediction analysis of data and forecasting of diagnoses.
Education
Higher education in Computer Science
Agency
400+
GMT-11
Remote
Want to hire this engineer?
Check if Viacheslav is available