Prototypes
A collection of AI and machine learning prototypes I've worked on, from research implementations to production systems.
AI-powered bot for detecting and moderating toxic messages in Telegram chats with automatic sanctions. Built using TensorFlow and Keras for real-time text classification.
Advanced customer segmentation system using K-Prototype algorithm to optimize banking marketing strategies. Handles mixed data types and provides actionable insights.
Machine learning models to predict customer churn and recommend retention strategies for telecom companies. Includes feature engineering and model interpretability.
Automated Captcha Reader
Intelligent script for automated captcha recognition and solving using computer vision techniques. Combines OCR with deep learning for high accuracy.
RAG System for Document Q&A
Production-ready Retrieval-Augmented Generation system for document question-answering. Supports multiple document formats with semantic search capabilities.
GPU Model Deployment Pipeline
Scalable pipeline for deploying machine learning models on GPU infrastructure with automatic scaling and monitoring. Handles real-time inference workloads.
Interested in collaborating?
I'm always open to discussing new prototypes and opportunities in AI and machine learning.
Get in Touch