Ivan Shelonik
About Ivan Shelonik
Ivan Shelonik is an Expert Machine Learning Engineer specializing in Generative AI, currently employed at Ciklum in Odessa, Ukraine. He holds a Master's degree in Processing of Big Data and has extensive experience in machine learning, computer vision, and developing applications using large language models.
Work at Ciklum
Currently, Ivan Shelonik serves as an Expert Machine Learning Engineer specializing in Generative AI at Ciklum, a role he has held since 2023 in Odessa, Ukraine. Prior to this position, he worked as a Senior Machine Learning Engineer at the same company from 2022 to 2023. In his current role, he has contributed to various projects, including the deployment of a Generative AI application and a large language model (LLM) on AWS EC2, utilizing AWS Cloud Formation and Docker Compose.
Education and Expertise
Ivan Shelonik holds a Master's degree in Processing of Big Data from the Belarusian State University of Informatics and Radioelectronics, which he completed in 2017. He also earned a Bachelor's degree in Biomedical/Medical Engineering from the same institution, graduating in 2016. His educational background provides a strong foundation for his expertise in machine learning and data processing.
Background
Before joining Ciklum, Ivan worked at BIQUANTS as a Machine Learning Engineer from 2019 to 2022 and as a Computer Vision Engineer from 2018 to 2019, both roles based in Minsk, Belarus. His experience in these positions has equipped him with a diverse skill set in machine learning, computer vision, and data engineering.
Achievements
Ivan has successfully developed a Museum Chat Assistant that features style and behavior adaptation, achieving significant reductions in average tokens per query and latency. He has also quantized and fine-tuned the LLama-2-13B LLM, integrating retrieval-augmented generation (RAG) for time-sensitive content. Additionally, he created a custom LLM Chat Agent for LangChain, which serves approximately 5000 users at a US Venture Firm.
Technical Skills and Projects
Ivan Shelonik has demonstrated proficiency in optimizing data pipelines, focusing on feature generation with large datasets stored in Iceberg and accessed through Trino using SQLAlchemy. His technical skills include Docker, AWS services, and machine learning frameworks, which he has applied in various projects to enhance application performance and user experience.