Erik Nikoghosyan

[email protected] • +374-77-73-78-74 • Yerevan
LinkedIn: www.linkedin.com/in/erik-nikoghosyan • GitHub: https://github.com/ErikNikoghosyan

PROFESSIONAL SUMMARY

Aspiring Machine Learning Engineer and Data Scientist with hands-on experience in developing and deploying diverse ML models, including Transformers, Autoencoders, and CNNs, gained through an R&D internship at Deep Origin. Proven ability to apply classical and advanced machine learning techniques—such as SVM, ensemble learning, and clustering algorithms—to real-world problems like protein analysis and sentiment classification. Skilled in Python, PyTorch, and data manipulation libraries (NumPy, Pandas), with a strong foundation in software tools including Git and Docker. Currently advancing expertise through formal education at Picsart Academy and YSU, eager to contribute innovative solutions and collaborate effectively in dynamic teams.

WORK EXPERIENCE

R&D intern
06/2025 - 09/2025
Deep Origin company
• Developed a handwritten Transformer model from scratch to solve the Netflix review classification task, deepening understanding of Transformer architectures and improving model interpretability
• Experimented with and implemented multiple optimization techniques, reducing model training time by 20% while maintaining accuracy, thereby enhancing computational efficiency
• Collaborated with senior researchers and cross-functional teams in an agile environment to co-develop and refine diverse AI models including autoencoders, CNNs, and neural networks, boosting data analysis accuracy by 15% during a 3-month R&D internship
• Presented technical insights and model performance results to the R&D team, influencing strategic decisions and promoting continuous innovation in AI solutions

EDUCATION

ML Engineer/Data Science Specialist
10/2024
Picsart Academy
Informatics and Applied Mathematics faculty
01/2023
Yerevan State Univerity , Yerevan

SKILLS

Technical Skills: Python, PyTorch, NumPy, Pandas, Machine Learning, Deep Learning, Neural Networks, Convolutional Neural Networks, Transformers, Autoencoders, Support Vector Machines, Decision Trees, Ensemble Learning, Boosting, K-Nearest Neighbors, DBSCAN, Principal Component Analysis, Singular Value Decomposition, Optimization, Statistics, Time Series Analysis, Calculus, Linear Algebra, Probability, Mathematical Analysis
Soft Skills: Teamwork, Problem Solving, Creativity, Adaptability, Communication, Self-Learning
Tools: Git, Docker, Jupyter Notebook, Kaggle, Virtual Environments, Server Management
Other: English (B2/C1), Russian (Intermediate), Armenian (Native), Machine Learning Research, Data Science Education, Picsart Academy, YSU Informatics

PROJECTS

Self Projects
Technologies: Python, NumPy, Pandas, Scikit-learn, Matplotlib, Jupyter Notebook, Custom MLP Neural Networks, Classical Machine Learning Algorithms, Data Preprocessing, Exploratory Data Analysis (EDA), Hyperparameter Tuning, Feature Engineering, Model Validation, Evaluation Metrics Selection, Focal Loss, Imputation Techniques, Predictive Modeling for Missing Data, Synthetic Data Generation, Noise-Resilient Classification Techniques
• Addressed challenges of imbalanced datasets by implementing advanced techniques such as Focal Loss, improving model sensitivity and overall classification performance by over 12% in real-world business scenarios
• Developed robust data imputation strategies for missing values by leveraging predictive modeling on correlated features, enhancing dataset completeness and increasing model accuracy by 8%
• Engineered and optimized a custom Multi-Layer Perceptron (MLP) from scratch for MNIST digit classification, achieving 96% accuracy through iterative architecture refinement and hyperparameter tuning using Python and NumPy
• Conducted experimentation on synthetic noisy datasets, including the "Noisy Flower" classification task, achieving above 88% accuracy by designing noise-resilient feature extraction and classification pipelines
• Led comprehensive data science workflows including Exploratory Data Analysis (EDA), feature engineering, and rigorous model validation, resulting in a 10% accuracy increase on Kaggle competitions involving house price prediction and personality trait classification
• Selected and aligned evaluation metrics such as accuracy, precision, recall, and RMSE with business objectives to ensure meaningful model assessment and deployment readiness

ACHIEVEMENTS

• Our team won 2nd place at the “LLM 4 ETL Hackathon” organized by MPP Insights in collaboration with Globbing, we worked on leveraging LLMs for ETL workflows — analyzing real customer datasets, detecting sentiment and intent, and generating actionable insights for business improvement.

Similar Resumes

MUHAMMAD ASIF High School Teacher House #38 Cnanal Breez Ali Town Raiwind Road Lahore, Punjab, Pakistan ASHOK Yennam, Venkata Krishna Reddy Project Manager Aachen,DE JASMEET ARORA SEO Intern Noida ANYI LISETH HOYOS TAPIERO Bogotá D.C. · Colombia