Tair Imankulov

Machine Learning Engineer and Computer Vision specialist with an MSc in Computer Science — I build end-to-end ML systems that ship, from research prototypes to real-world applications. My work spans two areas. On the deep learning side, I design multimodal systems for human activity recognition (94% accuracy) and emotion detection, with a focus on lightweight, efficient architectures — including a model running at just 0.5 GFLOPs for real-time inference on edge devices. On the generative AI side, I've built LLM-powered applications, including a multilingual (Kazakh / Russian / English) medical assistant using retrieval-augmented generation (RAG), and a real-time collaborative AI agent. Toolbox: Python, PyTorch, TensorFlow, Keras, OpenCV, LLMs & RAG, scikit-learn, Docker, Microsoft Azure (certified), PostgreSQL. I enjoy designing architectures, squeezing performance out of models, and applying AI to problems that matter. I'm currently open to remote ML/AI engineering roles, or on-site opportunities in Kazakhstan — I'd love to join a team building something ambitious.