aydin abedinia

I am a Senior Machine Learning Researcher and Distributed Systems Engineer with over eight years of experience creating AI solutions that deliver substantial business value. My work focuses on the intersection of academic research and practical industry application, where I've generated significant business impact through models for rider-driver matching, adaptive pricing, and real-time fraud detection. I thrive on architecting these systems for massive-scale environments, having deployed solutions serving over 70 million users and processing more than 5.9 million transactions daily at Snapp!, the Middle East's leading mobility platform.

Beyond my professional roles, I am deeply committed to advancing the machine learning field through peer-reviewed research and open-source contributions. My academic work focuses primarily on semi-supervised learning, leading to publications such as the "Semi-CART Algorithm" in the Springer journal and a presentation on distance-based sample weights at the IEEE ICMLT conference. To make these innovations accessible to the broader community, I've developed and actively maintain several open-source libraries, including production-ready implementations of SemiCART and a high-performance Decision Tree library written in Rust.

My technical expertise covers the full AI lifecycle, from algorithm development in Python, Go, and Rust to advanced modeling with frameworks like PyTorch and TensorFlow. I specialize in building and deploying scalable, containerized microservices using modern MLOps practices and tools like Kubernetes, Docker, and Apache Spark. Grounded in a strong academic foundation and driven by a passion for solving complex, large-scale data problems, I am always eager to translate novel research concepts into impactful, real-world solutions.