Bio
Nilakash Das is a final year Ph.D. researcher in the field of clinical data science. His work lies at the intersection of artificial intelligence and respiratory medicine. Recently, he was awarded the prestigious Young Scientist Scholarship by the European Respiratory Society.
Currently, he works at the laboratory of respiratory medicine and thoracic surgery at KU Leuven, Belgium. He also collaborates extensively with Artiq, a spin-off company of their laboratory, in translating AI research into clinical practice.
Summary
In this episode, Nilakash Das delves into how he became a clinical data scientist, his work on using deep learning for Spirometry and offers a workable approach for students to break into AI for healthcare.
Key Questions
9:02.48: ..you had all these aspirations and you came into this (Healthcare + AI) – so how much of this are you able to do in your day-to-day work as a clinical data scientist?
14:14.20: Particularly now that you are very engaged with this field, especially in the case of respiratory medicine, what is the challenge that you see here?
26:42.19: Before we get further…could you just explain in brief what is Spirometry, what is it used for?
36:54.63: So how are you guys thinking of bringing this work into the clinical workflow so that it’s more mainstream?
Quotable Quotes
9:57.18: Where we try to solve or target a specific clinical outcome using a model – then once the model is developed, we have to design clinical studies or do clinical development to prove the efficacy of the AI model.
23:09.47: But definitely there’s a huge disparity into the amount of research that we are producing and the number of things that are actually making it to real practice.
25:31.23: So, as the AI algorithms become more powerful and as they become autonomous who is going to be liable for any incorrect clinical decision-making?
35:14.45: ..the general practitioners are common people like us and they often enrol in a Spirometry training course. You have to shell out a lot of money. To let your nurse enrol in a training program you have to shell out a lot of money and naturally this is not easily affordable to most people.
41:13.47: Because almost all the jobs in healthcare or even in the biotech sector require proven experience with biological or human data and that is something which the courses or boot camps or even doing Kaggle competitions does not help.
42:40.13: And one of the ways that I often suggest is after you have done a machine learning course, you should try to reach out to a biologist or a doctor in your network because these folks often have access to real world clinical or biological data.
Mentions of Books, Movies, Podcasts, Websites, Events, Places, People
3:38.74 Andrew Ng’s Stanford’s lectures on AI posted on YouTube.
3:40:15 AlexNet, a CNN that outperformed all computer vision algorithms in the ImageNet challenge in 2012.
38:17.49 European Respiratory Society.
Connect
LinkedIn: Nilakash Das
The You+AI Vodcast
A companion video segment full of fun and candid moments.
Watch it here!