Petr

PETR YERMAK
Junior Machine Learning Engineer
Spain

[відкрити контакти](див. вище в блоці «контактна інформація»)

GitHub: github.com/YermakPetr

[відкрити контакти](див. вище в блоці «контактна інформація»)

PROFESSIONAL SUMMARY

Junior Machine Learning Engineer with hands-on experience developing machine learning and computer vision projects
using Python, Scikit-learn, OpenCV, and MediaPipe. Skilled in data preprocessing, feature engineering, model evaluation, and
building end-to-end ML pipelines. Continuously improving practical machine learning skills through personal projects and
seeking a Junior Machine Learning Engineer position where I can apply my technical skills, contribute to real-world projects,
and continue developing as an ML engineer.

TECHNICAL SKILLS

Programming: Python

Machine Learning: Scikit-learn

Computer Vision: OpenCV, MediaPipe

Data Analysis: Pandas, NumPy

ML Concepts: Data Preprocessing, Feature Engineering, Model Evaluation

Tools: Git, GitHub, Jupyter Notebook, VS Code

SELECTED PROJECTS

Slot Style Classifier

Computer Vision | Unsupervised Learning

● Developed a computer vision pipeline to analyze slot games using both static screenshots and gameplay videos.
● Extracted spatial and temporal features with OpenCV and applied unsupervised learning (K-Means) to discover
gameplay patterns.
● Compared static image classification with dynamic video analysis and demonstrated the limitations of single-frame
approaches.
● Built data preprocessing, feature extraction, visualization, and clustering pipelines using Python and Scikit-learn.
Technologies: Python, OpenCV, NumPy, Pandas, Scikit-learn, SciPy, Matplotlib

GitHub: github.com/YermakPetr/slot-style-classifier

ErgoVision AI

Computer Vision | Pose Estimation | Video Analytics

● Developed a computer vision pipeline to analyze long videos using MediaPipe pose and face landmarks.
● Extracted posture and facial activity features, aggregated measurements over time, and detected unusual events
using z-score anomaly detection.
● Automated the generation of ranked video clips and reports, reducing manual review of 90-minute recordings to a
small set of relevant segments.
● Built an end-to-end pipeline including feature extraction, temporal aggregation, anomaly detection, visualization,
and FFmpeg-based video export.

Technologies: Python, MediaPipe, OpenCV, Pandas, Matplotlib, FFmpeg

GitHub: github.com/YermakPetr/ErgoVision-AI

EDUCATION

Odesa National Polytechnic University

Specialist Degree in Radio Engineering

Bachelor's Degree in Radio Engineering

LANGUAGES

Russian — Native

Ukrainian — Native

English — Technical Reading

Czech — Basic