A recent interview between Rebecca Schankerman, Ambra Health Field Marketing Coordinator, and Alexey Kachalov, Ambra Health Data Scientist, about his passion for machine learning and engineering.
Rebecca: Can you talk about your time working at Ambra and how long you have been with the company?
Alexey: I have been working at Ambra since September 2019. Working at Ambra has allowed me to develop so many skills, and it has been fantastic for me to develop different ML models. It’s great to know that what I work on brings direct value to healthcare.
Rebecca: What does a typical day look like for you and the Engineering team?
Alexey: We have LeanKit task tracking system where all personal and team tasks are available. My typical day consists of coding, data analysis, and discussion of ongoing model development.
Rebecca: What do Ambra engineers do for fun? Do you participate in coding challenges or contests?
Alexey: Yes, I participated in the RSNA 2020 STR Pulmonary Embolism Detection Kaggle competition. The aim of the competition was to detect and classify pulmonary embolism cases in chest CT studies with contrast. I have developed 2 machine learning models for this competition – the first model identifies pulmonary embolisms on the DICOM level and the second model classifies left sided, right sided, and central pulmonary embolisms. Other parameters such as the RV/LV ratio present in the study is >= 1, acute and/or chronic type of PE on the study level. In total, about 800 teams from all over the world took part in the competition. My final solution took 29th place and got the Kaggle Silver medal.
Rebecca: Are there unique challenges in writing code for a healthcare company?
Alexey: There are a lot of ongoing challenges in the medical domain on the Kaggle: HuBMAP – Identify glomeruli in human kidney tissue images, RANZCR CLiP – Classify the presence and correct placement of tubes on chest x-rays to save lives, VinBigData Chest X-ray – Localize and classify thoracic abnormalities from chest radiographs.
Rebecca: What projects are you most excited about currently & in the next year?
Alexey: Implementation of a billing code lookup model for a few new customers, development of DWF model for text segmentation on PHI and non-PHI parts, and development of a fast and reliable OCR text recognition model.
Rebecca: The Ambra Health team is currently working remotely. How is quarantine treating you?
Alexey: I believe that quarantine is a temporary phenomenon, but it should teach us to be ready to work in a distributed manner.
Alexey Kachalov is a graduate of Tomsk State University of Control Systems and Radioelectronics with an honors degree in Information Technology. Alexey has a strong passion for machine learning and how it plays a role within the healthcare industry. As a Data Scientist at Ambra Health, Alexey develops new tools for Ambra’s AI services.