AI based applications in healthcare have made great strides especially in the field of computational pathology, thus providing high quality clinical services !
Harshita Sharma, our next pathbreaker, Senior Researcher at Microsoft Research (Cambridge, UK), works in the Medical Imaging team, exploring machine learning solutions aimed towards improving patient outcomes and clinical workflows.
Harshita talks to Shyam Krishnamurthy from The Interview Portal about her PhD research focused on analysing whole slide images of gastric cancer using machine learning methods and deep learning architectures.
For students, scientific research might seem daunting at first, but if you have made up your mind, no problem is big enough to unnerve you !
Harshita, can you take us through your background?
I was born in Delhi. My parents are medical doctors – they carved their own career paths and inspired me to dream big and work hard to achieve my goals. Although I grew up in an academically inclined family focused on studies, I balanced my student life with hobbies such as writing poetry, singing, playing the harmonium, and drawing.
In school, I enjoyed studying physics, math, and biology, and wondered if there was a career that could perhaps combine these streams? Moreover, I felt motivated to work on solving problems that can create an impact in society and can directly benefit humanity. This perhaps explains my unusual career path. I am a computer scientist and engineer developing innovative solutions aimed at improving healthcare. I have an interdisciplinary background with experience (industry, research, teaching) and qualification in computer science, biomedical, and electrical engineering.
What did you do for graduation/post graduation?
In 2010, I obtained a bachelor’s degree (B.Tech.) in electronics and communication engineering from the Indira Gandhi Delhi Technical University of Women. In 2012, I obtained a master’s degree (M.Tech.) in electrical engineering with focus on signal processing, from the Indian Institute of Technology Roorkee. In 2011-12, I was awarded an exchange research scholarship at the Technical University of Berlin under the DAAD-IIT Masters Sandwich Programme. In 2013, I was awarded the DAAD full-time PhD scholarship to pursue my doctoral studies. In 2017, I received my PhD degree (doctorate in engineering “Dr.-Ing.”) in computer science from the Technical University of Berlin. In 2017-21, I was a postdoctoral researcher in biomedical engineering at the University of Oxford.
What made you choose such an offbeat, unconventional and uncommon career?
Coming from a medical family background and studying math with biology in school, I hadn’t received any training or coaching in engineering / computer science before my bachelor’s degree. Being competitively selected at a public engineering college and enrolling for my bachelor’s degree were probably the first defining moment for my career path as the learnings, and experiences that I gained there built firm foundations for my future journey and accomplishments. Moreover, since my formative years of training, I have been attracted to interdisciplinary research combining engineering and technology with medicine and healthcare.
Towards the end of my bachelor’s degree, I was selected on-campus by 3 tech MNCs and for a scientist position at ISRO (cleared its entrance exam with a rank in 20s). However, I wanted to explore academic and research prospects further, and so I decided to pursue a master’s (M.Tech.) at IIT Roorkee, after qualifying the GATE exam. My decision may not have been the most conventional one, but I feel grateful to be supported by visionaries like my parents, grandparents, and mentors, who were encouraging and appreciative of my desire to pursue higher studies. The outcomes were more rewarding than we imagined. During my master’s, I traveled to Germany on a research exchange scholarship, then I completed my PhD on a full-time scholarship, did a postdoc at the University of Oxford, and currently work as a senior researcher at Microsoft Research in Cambridge. Throughout my career so far, I feel extremely fortunate to have met and worked with some of the brightest and most talented people in the world.
How did you plan the steps to get into the career you wanted? Or how did you make a transition to a new career? Tell us about your career path
I have always wanted to apply my knowledge and skills towards technical and societal impact, so I pursued interdisciplinary projects combining engineering with healthcare, though persistently focusing on medical imaging and biomedical data science. I worked on multiple research projects on topics like medical image processing and analysis, machine learning applied to medical imaging and biomedical data, and speech analysis during my bachelor’s and master’s. For instance, in my third year B.Tech., I did a research internship at DRDO where I developed a speaker recognition system. During my bachelor’s, I presented three research papers at national conferences, including a conference at IIT Roorkee, which was my first breakthrough in scientific research. I am grateful to my mentors during my bachelor’s degree as they played a substantial role in introducing me to scientific research practices such as paper writing and presentation skills.
During my M.Tech. at IIT, I was selected for full-time research in Germany via the IIT-German Academic Exchange Service (DAAD) Masters Sandwich Scholarship Programme, where I worked in the Computer Vision research group at the Technical University of Berlin in a joint research project with Charité University Hospital Berlin, to analyze breast cancer images in digital histopathology. This period of knowledge and cultural exchange was extremely rewarding and helped me grow further as a researcher. I received hands-on experience of cutting-edge research in the field, wrote my first journal publication, travelled across Europe, gained international exposure, and succeeded in finding avenues for further research collaborations. On my return to India, I became a Lecturer at Jaypee Institute of Information Technology in Delhi-NCR.
Soon after, I was awarded the DAAD full-time PhD scholarship to pursue my doctoral studies supervised by Prof. Olaf Hellwich in the Computer Vision research group at TU Berlin. The goal of my PhD was to analyze whole slide images of gastric cancer using machine learning methods, including graph-theoretic handcrafted features and deep learning architectures. I collaborated with researchers and clinicians at the Institute of Pathology, Charité University Hospital, Berlin and University Hospital Schleswig-Holstein, Kiel. My PhD was the next biggest defining moment in my professional and personal life, a period of immense growth and learning when I was deeply engaged in research activities, received several opportunities to present my work in conferences around the world, and published research papers in peer-reviewed scientific journals and conference proceedings. In this time, I gained specific technical and domain knowledge in fields such as machine learning for health, biomedical image analysis, computational pathology, deep learning, and big data science. After 3.5 years, I was awarded a doctorate in engineering “Dr.-Ing”. I continue to contribute to the TU Berlin alumni activities in the UK as the international alumni contact, and co-founder and manager of “TU Berlin Alumni Club UK” on LinkedIn.
Less than a week after my PhD viva, I joined the Institute of Biomedical Engineering at the University of Oxford as a postdoctoral researcher. This was an incredible opportunity where I analyzed rich real-world fetal ultrasound data to develop novel computer-aided methods. I did individual research and collaborated with colleagues in the research group and clinicians at the Oxford University Hospitals (NHS). Additionally, I was co-supervising PhD students and was involved in university teaching, organizing, and volunteering activities. I developed my teaching portfolio at the University of Oxford through lectures, tutorials, and labs, and received the SEDA PDF Supporting Learning Award, a portable accredited qualification for Higher Education teaching in the UK.
In general, I have always been inclined to work on research projects with real-world applications. Specifically, during my postdoc at the University of Oxford, I worked in the ERC Project PULSE (Perception Ultrasound by Learning Sonographic Experience https://eng.ox.ac.uk/pulse/) with the aim to develop multimodal methods to model how a sonographer makes decisions from ultrasound images and build assistive interpretation tools in routine fetal ultrasound. For instance, I developed methods for full-length scan video description, clinical workflow analysis, and pupillometry analysis for fetal ultrasound, the first in Sonography Data Science with potential real-world applications of understanding sonographer workflow and assessing their skills and cognitive workload during routine scanning. In this time, my research skills became much broader and deeper, as I not only refined my existing interests in machine learning and biomedical image analysis, but also extended these to novel areas such as ultrasound imaging, video analysis, natural language processing, eye-tracking, and multimodal data analysis. My postdoc mentor at University of Oxford, Prof. Alison Noble, is a role model for several women in STEM including me, and I am very fortunate to receive her leadership and guidance.
After my postdoc, I joined Microsoft Research Cambridge in the Health Intelligence research theme (https://www.microsoft.com/en-us/research/theme/health-intelligence/), with the goal to explore and develop machine learning applications for improving clinical pathways and revolutionizing healthcare.
How did you get your first break?
I consider my first big break in research with the DAAD-IIT Master Sandwich Program Scholarship, a full-time research scholarship during M.Tech second year. In 2011-12, I received the scholarship for 9 months, through which I traveled to Germany and completed my master’s thesis at the Technical University of Berlin. I heard about the scholarship through my classmates at IIT. I did some more research and identified institutions in Germany working on my areas of interest. I requested my IIT supervisor to contact research group heads via email with my CV and proposal, and he received a positive reply from our first attempt (which was also my first preference). The outcome of this work was a successful master’s dissertation at TU Berlin and my first journal paper, where I analyzed histological breast cancer images for content-based image retrieval using graph theoretic description and matching methods. The collaboration between my IIT and TU Berlin supervisors grew further in the next years, as I saw several successors from IIT pursuing research through this new collaboration channel, which makes me feel so gratified and rewarded. Based on my research and academic accomplishments, I was selected for a full-time DAAD PhD scholarship next year.
During my higher education, PhD, and postdoc, I have worked on extremely interesting machine learning and image analysis methods to address real-world problems in healthcare. Also, I have actively followed and admired the vision, ambition, and activities of Microsoft towards impactful research and solutions in healthcare. Hence, I was motivated to apply at Microsoft after my postdoc. Working in the industry would be novel and exciting for me, with immense opportunity to learn and develop new skills and expertise and apply these for real-world applications and challenges.
What were some of the challenges you faced ? How did you address them?
Initially, a big challenge was moving from India to Germany, after spending more than 20 years in Delhi with my family. Living and working in a new environment all by myself was transformational and a big learning curve (for e.g., culture, language, food) and it gave me a great sense of independence and responsibility. I am thankful to DAAD for organizing cultural activities and German language courses that gave me the opportunity to socially interact and network locally in Berlin. Solo travelling to new countries for conferences and meetings seemed daunting in the beginning but became very enjoyable over time. I think I relived some experiences moving from Germany to the United Kingdom for my postdoc, but it was much smoother and rewarding this time, perhaps because of my previous learnings and experience.
Where do you work now? What problems do you solve?
I am a senior researcher at Microsoft Research in Cambridge, United Kingdom.
I am working in the Medical Imaging team, where I explore machine learning solutions aimed towards improving patient outcomes and clinical workflows. My research interests include machine learning, image analysis, computer vision, medical imaging, and multimodal clinical data analysis for healthcare. I want to create large-scale technical and societal impact with my work, and the role provides me with this opportunity by contributing towards cutting-edge research and innovations in healthcare. For instance, the developed machine learning solutions at Microsoft Research are being used in hospitals (e.g., https://www.microsoft.com/en-us/research/project/medical-image-analysis/) and by researchers (e.g., https://www.microsoft.com/en-us/research/project/project-innereye-open-source-software-for-medical-imaging-ai/ ).
How does your work benefit society?
In my field of expertise, the last few years have witnessed massive interest from different sectors. Artificial intelligence and machine learning continue to make a greater impact in our daily lives, and healthcare has also benefited from many such technological advancements.
Tell us an example of a specific memorable work you did that is very close to you!
My PhD research is very special to me. It received a lot of attention from the scientific community, for instance, best paper awards, mentions in articles and blogs, and many citations. My work was probably the first to try solving the problem of automatically identifying cancer types based on genetic alterations, from the more commonly and routinely used H&E stain in histopathology whole slide images, in place of special (IHC) stains, using (deep) machine learning methods. The work is very promising as it could benefit healthcare where special staining or genetic tests are expensive and/or time-consuming.
Your advice to students based on your experience?
I feel privileged to realize that I can potentially be a role model to the next generations, especially to the women in STEM. My sincere advice to students would be to persist, persevere, and focus on your goals and it is never too late to consider a career in scientific research. There are many attractive research pathways that one may choose from, but it is advisable that the choice is based on one’s interests. It is always a good idea to start with small projects and read the literature to understand your unique research interests.
I have seen difficulties and obstacles in my journey, but my determination has empowered me to overcome these and carve my unique career path. I have changed my physical location over the past years and evolved as a person, enhanced my skills, and enriched my experiences whilst persistently focusing on my contribution to science and technology. In the positions I have been in so far, I have loved my day-to-day job and will continue to make this my priority. Outside of work, I like spending time with family, playing, cooking meals, watching movies, and going for walks.
I am a lifelong learner, so I will continue to learn, improve my skills, and expand my knowledge in the future as well – that’s the only plan for now 😊
Further details about me can be found at