🔍 Computer Vision & Deep Learning: Developed and fine-tuned a custom CNN-LSTM architecture for video classification. 🧠 Model Design & Optimization: Used pre-trained DenseNet-121 for spatial feature extraction and ConvLSTM2D to capture temporal dynamics across frames. 📈 Data Pipeline Engineering: Built a custom FrameGenerator to preprocess and batch video datasets (OpenCV + TensorFlow/Keras). 🎞️ Multi-Dataset Evaluation: Validated model across 4 real-world datasets (Hockey, Movies, Violent Flow, RLVS) showing high accuracy with low computational overhead. ⚙️ Tools & Technologies: Python, TensorFlow, Keras, OpenCV, NumPy, scikit-learn 🧪 Research Methodology: Framed anomaly detection as a weakly supervised learning problem; conducted literature review on deep anomaly detection and scoring methods. 🚀 Real-Time Readiness: Focused on optimizing model for deployment in real-time surveillance scenarios with constrained resources.

-MySQL Database Design & Modeling / Data Extraction & Querying -QlikSense Dashboard Data Analysis & Reporting -Worked with Kafka producers and consumers to stream data between microservices -Wrote Python scripts for data extraction, transformation, and loading

- Консультирование клиентов по вопросам выбора системы страхования - Выявление потребности клиента в продукции и услугах компании - Ведение клиентской базы - Заключение договоров страхования - Поиск и привлечение новых клиентов - Составление банковских документов, платежных поручений, требований и ордеров