Computer VisionDec 15
Building Real-Time Movement Recognition with OpenCV
How I developed the computer vision system for Reviva using smartphone cameras and PyTorch to provide instant feedback for rehabilitation exercises.
Read article
Arseniy
Kirykouski
Machine Learning · Data Science
Milan, Italy
Open to Machine Learning & Data Science roles
About
AI undergraduate with hands-on experience in machine learning, NLP, computer vision, and data analytics. Proficient in Python (pandas, NumPy, scikit-learn, PyTorch, OpenCV) and skilled at turning complex data into actionable insights.
Passionate about applying AI to real-world business challenges and a consistent winner in international hackathons.
🏆
3
Hackathon Awards
🎯
97%
Model Accuracy
📚
4
Published Projects
📊
27/30
GPA in ML Courses
Education
B.S. Artificial Intelligence
University of Pavia, Milan, Bicocca
GPA
27/30
Machine Learning: 27/30
Achievements
Hackathon Winner
EIT Health, Bologna StartUpDays
Experience
AI Research Challenge Winner
Sopra Steria
Contract
2024-09 – Present•Remote
Leading development of Reviva, an AI-powered post-stroke rehabilitation application. Architecting computer vision systems for real-time movement recognition and building transformer-based NLP solutions for medical diagnosis.
- ▪Won Sopra Steria International Student Challenge with innovative healthcare AI solution
- ▪Achieved 97% accuracy on medical diagnosis classification with fine-tuned transformers
- ▪Implemented real-time computer vision system processing smartphone video at production scale
- ▪Presented work to 150+ industry leaders at international hackathon event
PythonPyTorchOpenCVHugging FaceBERTClinicalBERT
Research Assistant
University of Pavia / Milan
Research
2023-09 – 2024-06•Milan, Italy
Conducted independent research on advanced machine learning and NLP applications. Developed multiple projects combining computer vision, transformers, and statistical models.
- ▪Developed 3 industry-grade ML projects with published results and documentation
- ▪Fine-tuned medical NLP models (BERT, ClinicalBERT, Flan-T5) on healthcare datasets
- ▪Implemented comprehensive data pipelines for preprocessing and feature engineering
- ▪Published blog posts with 100+ technical details on implementation approaches
PythonPyTorchscikit-learnXGBoostpandasHugging Face
Currently Working On
Expanding Reviva, the AI-powered post-stroke rehabilitation application, with enhanced movement recognition algorithms and personalized treatment plans.
Exploring advanced transformer architectures for medical NLP applications and continuing research in computer vision for healthcare.
Skills
PythonPyTorchOpenCVHugging Facescikit-learnPandasNumPyXGBoost
Projects
Click to expand and see details
Reviva
2024
Winner - Sopra Steria International Student Challenge
AI-powered application for personalized post-stroke rehabilitation. Implemented patient movement recognition using OpenCV and smartphone camera input. Trained a model to provide real-time exercise feedback based solely on visual data.
Real-time CV feedbackHealthcare use-caseHackathon winner
PythonOpenCVPyTorchHugging Facescikit-learn
SMP2DIAG
2024
NLP classifier predicting 22 medical diagnoses from patient symptom descriptions (GretelAI dataset, 1,065 samples). Fine-tuned transformer models: BERT/ClinicalBERT (97% acc, F1=0.97) and Flan-T5 (94% acc), outperforming classical baselines.
97% accuracyClinicalBERT22 diagnoses
PythonPyTorchHugging Facescikit-learn
Bike Sharing Demand Prediction
2024
Regression models to predict bike sharing demand from weather, time, and calendar features. Implemented preprocessing (scaling, one-hot encoding, PCA/LDA) and trained models including Linear/Ridge/Lasso/ElasticNet, Random Forest, and XGBoost.
Feature engineeringEnsemblesDemand forecasting
Pythonscikit-learnXGBoostpandasnumpy
Open Source & Projects
Featured
Reviva
by se4wer
AI-powered post-stroke rehabilitation application using computer vision and NLP. Open-sourced to help researchers and developers build healthcare AI solutions.
Creator & Maintainer - Computer vision pipeline, model training, API design
View on GitHub →
SMP2DIAG
by se4wer
Medical NLP classifier predicting diagnoses from symptom descriptions. Demonstrates transformer fine-tuning and medical AI best practices.
Creator & Maintainer - Data processing, model implementation, evaluation framework
View on GitHub →
Hackathon Achievements
EIT Health i-Days
3rd PlaceMadrid · 2025 · 850 participants
REVIVA - AI-powered app for post-stroke rehabilitation with personalized plans and real-time feedback
Bologna StartUpDays
WinnerBologna · 2024
Winner in "Biomedicine & AI" track
EIT Health i-Days
3rd PlaceItaly · 2024 · 250 participants
Represented Italy at the European Finals; received mentorship from healthcare and innovation experts
Certifications & Awards
Machine Learning Specialization
Andrew Ng (Stanford)
2024-03
Supervised LearningNeural NetworksUnsupervised Learning
Advanced Python for Data Science
University of Pavia
2023-12
Advanced PythonData AnalysisNumPyPandas
1st Place - EIT Health i-Days
EIT Digital
2025-01
Healthcare InnovationAI Application
3rd Place - Bologna StartUpDays
Bologna StartUpDays
2024-05
EntrepreneurshipProduct Development
3rd Place - EIT Health i-Days Italy
EIT Digital
2024-04
Healthcare AIInnovation
Education
B.S. in Artificial Intelligence
Joint Degree: University of Pavia | University of Milan | University of Milan Bicocca
Milan, Italy
GPA: 27/30
Machine Learning & Deep Learning27/30
Advanced Statistics & Probability28/30
International Baccalaureate Diploma Programme
New School, International School of Georgia
Tbilisi, Georgia
GPA: 42/45 (top 3% of students)
Mathematics & Physics focus
Languages
English
FluentC1
Italian
IntermediateB1
Russian
NativeNative
Recommendations
MB
Dr. Marco Baroni
Professor of Machine Learning
University of Pavia•Academic Mentor
"Arseniy demonstrates exceptional depth in machine learning theory and practical implementation. His work on medical NLP transformers shows mature understanding of both technical and domain-specific challenges. Highly recommended for advanced AI roles."
2024-12
ER
Elena Rossi
Judge, Sopra Steria Challenge
Sopra Steria•Hackathon Evaluator
"Arseniy's Reviva application was outstanding. The computer vision implementation was technically sound, the healthcare domain expertise was evident, and the presentation was professional. Winner-worthy work across all dimensions."
2024-09
Thoughts & Blog
Insights from projects and experiments
NLPNov 20
Fine-Tuning Transformers for Medical Diagnosis
A deep dive into achieving 97% accuracy with BERT and ClinicalBERT for symptom-to-diagnosis classification, and the challenges of working with medical datasets.
Read article
ProjectsOct 10
From Hackathon to Production: The Reviva Journey
Reflecting on winning the Sopra Steria International Student Challenge and the lessons learned in building an AI application for healthcare.
Read article
Quick Interest Check
Let me know what opportunities interest you. I'll get back to you within 24 hours.
Frequently Asked Questions
Common questions from recruiters about availability, experience, and working together.
Availability
Yes! I'm actively open to Machine Learning Engineer, Data Science, and AI Research roles. I'm graduating in June 2025 and available for full-time positions from that date. Open to internships and contracts in the interim.
Work
Location
Technical
General
Testimonials
"Arseniy demonstrated exceptional technical skills and innovative thinking during the hackathon. His ability to apply AI to real healthcare challenges is impressive."
Hackathon Judge
EIT Health i-Days