Ivan Drokin Email & Phone Number

Budva, Montenegro

Designed research projects, including methods and models selection, evaluation processes, experiments plans, exploratory data analysis, models finetune and improvements. Developed CV models for industrial/customer electronics applications (see key projects below) Participated at delivery processes, including lab tests and demos. Public activity at professional conferences and meetups as well as popular science events. Key projects: Deep learning solution for objects localization at global scene using images from set of cameras.Deep learning-based odometry solution with visual, IMU and GPS data fusion for outdoor self-localization. Key results: 6 satisfied customers 20+ successfully implemented projects A colleague successfully defended his master's thesis under my supervision.

Designed, developed and evaluated pipelines for medical image analysis (Chest and Brain CT scans, Chest X-rays): instance and semantic segmentation, objects detection, classification. Worked with and tutored the Medical team to establish a clear data gathering and labeling process. Designed transparent and thorough models evaluation protocols including medical validation step. Tutored Junior/Middle Researchers. Leaded ML and Medical teams. Designed and implemented ML research process, and entire ML models lifecycles. Collaborated with DevTeam and DevOps Team to develop efficient production and deployment solution. Helped Business and Product Team to formulate new areas of research, discover new products and technologies.Presented the company, products and technologies at professional conferences and meetups Key results: Achieved SOTA results at lung nodule screening on Chest CT task (results were presented at MLMI MICCAI 2021). Created ML application pipelines for lung cancer screening on Chest CT. This solutions were successfully implemented in more than 50 pilots and commercial projectsPresented list of paper at topical ML and medicine conferences (MIDL 2018, MLMI MICCA 2021, AIST 2019-2020, AINL 2018) Raised Research team of 6 DL researches and 8 radiologists

Designed, developed and evaluated: Mathematical models for cell lines optimization Deep Learning models for small molecules properties prediction Drug portfolio augmentation system based on Deep Learning and Genetics Algorithms

Developed market information analysis and decision making system (trading robots) based on ML algorithms for time-series. Developed robust platform integrated with broker's API