Ruchika Verma: introducing industry to the benefits of machine learning

Ruchika Verma's career as a machine learning scientist is young but already impressive. Her work combines health data, machine learning and medical imaging to push forward the boundaries of precision medicine and cancer research. She's organized international competitions to find innovative methods of identifying, segmenting, and classifying tumour cells.

Furthermore, she was recently recognized with Case Western Reserve University's prestigious Outstanding Graduate Career Award for her work in using machine learning for treatment outcome and survival prediction in glioblastoma (GBM), an aggressive form of cancer. Now, she’s set her sights on accelerating AI adoption on Amii’s industry, helping companies build their internal AI capabilities.

It’s a career that was started by an unlikely spark — a computerized game of Hangman.

Ruchika's path to becoming a machine learning researcher started during her undergrad, studying Electronics and Communication Engineering at Punjab Technical University in Jalandhar, India. She ended up taking a few classes that involved programming, where one of her first projects was to create a digital version of the classic letter-guessing game. Immediately, she was hooked. While it was a simple program, Ruchika says her Hangman game got her immediately hooked on the creative aspect of coding.

"When I first got these assignments, I felt so happy, so accomplished," she says. "With code, you can do something and bring your ideas to life."

That fascination with coding stayed with her as she continued her education, but it was joined with a new interest — medicine. Specifically, Ruchika says she was interested in pathology, the study and diagnosis of disease through the examination of surgically-removed tissue sections. After finishing her undergrad, Ruchika received a master's from the prestigious India Institute of Technology Guwahati with a specialization in statistical signal processing and machine learning. She also taught undergraduate courses on digital image and signal processing at the National Institute of Technology, Meghalaya — the Government of India’s institute of national importance.

During her masters, Ruchika found that she could combine her interests in programming and pathology, investigating whether machine learning could help physicians spot tumours with digital microscopes. However, what most attracted her to the work was the chance to do something that would significantly impact people's lives. For instance, her research included application of machine learning to better identify the molecular sub-types of breast cancer in patients.

A patient might receive a diagnosis of breast cancer, for example. But that cancer can have further subtypes that can drastically change their chances of survival and which treatment is the most effective for a patient. Mostly, this identification is made by a pathologist by visual inspection of tumour samples taken during a biopsy or surgery. But different locations on the same tumour might have multiple subtypes of cancer cells, which manual identification can easily miss.

Ruchika's work led to identifying multiple sub-types in a single patient using machine learning that could aid in suggesting personalized medicine because the detection of a secondary sub-type in a single patient may impact the prognosis and treatment of the cancer.

"With humans looking at the tumour, we generally only get to know the dominant subtype of a tumour. But with machine learning, we were able to find the co-existence of multiple sub-types that could change the treatment for a patient," she says.

She also won a travel award from the IEEE signal processing society to present her work on breast cancer at the IEEE International Conference on Image Processing (ICIP) in Phoenix, USA.

"I wanted to implement something that would have an impact. What is the point of doing something if it isn't adding value to people’s lives?"

Identifying cancer accurately

Ruchika then took the long journey to North America, where she pursued a PhD in Biomedical Engineering at Ohio's Case Western Reserve University. While at Case Western, her research work comprised developing new image-based markers for prognosis of overall survival as well as prediction of response to therapy in GBM tumours. It involved training machine-learning models to look for markers in MRI scans and pathology slides (tissue samples) from tumours that could help predict how a GBM patient would respond to chemo-radiation therapy. Ruchika graduated with eleven journal papers and a patent, as well as a number of conference papers and abstracts in leading scientific conferences. Her most recent research is published in high-impact journals such as IEEE Transactions on Medical Imaging, Clinical Cancer Research, Journal of Pathology, Medical Physics and has garnered significant interest from the clinical community. Case Western recognized her excellent research work by awarding Ruchika with the prestigious Outstanding Graduate Career Award.

During her graduate studies, she also co-organized two international competitions on nuclei segmentation. One on cancer nuclei segmentation and classification, MoNuSAC at the IEEE International Symposium on Biomedical Imaging (ISBI 2020), where 46,000 hand-annotated nuclear boundaries of tumour cells (epithelial and immune) were released to facilitate future research for automatic characterization of cancerous cells. She also co-organized another international competition on nuclei segmentation, MoNuSeg at MICCAI 2018, where she created a rich resource for future research in computational pathology.

During her PhD, Ruchika also conducted her research for colorectal tumour localization and survival prediction using artificial intelligence as a research intern, respectively, at the University of California and the University of Alberta.

A much larger impact

After finishing her PhD, Ruchika joined Amii's team in January 2022, originally as a machine learning intern. She partnered with Canada's National Research Council (NRC) to investigate whether machine learning could be applied to predict biomolecule abundance of Canadian protein crops. The objective of the project was to develop artificial intelligence (AI) and ML solutions to extract patterns from plants and to predict molecular phenotypes directly from their DNA sequences, which could help research to increase crops yield or nutritional content. While it might seem like a vast departure from her work with medical imaging, Ruchika says they have one important thing in common: the chance to impact people.

"I wanted to work on something where there could be real value to Canada. My previous work had an impact on people, but finding ways to extract meaningful patterns in Canadian protein crops can aid in growing food with better protein content and improve yield production, that could affect a much larger number of people," she says.

In April, Ruchika signed on full-time with Amii as an Associate Machine Learning Scientist with the Industry team. She now works with businesses and other organizations, offering her machine learning expertise to solve real-world problems. Working directly with industry has allowed her to broaden her understanding of the potential that machine learning offers: in addition to her work with the NRC, Ruchika’s projects have included using AI to suggest new battery designs and models that help flag racist comments on social media.

Her work with the Industry team also allowed her to pursue the passion that has so far fuelled her impressive career — work that positively impacts the world. It also brought her to a city she has fallen in love with, where she enjoys river valley hikes and occasional mountain excursions with her husband.

"Coming here from India was a long journey. Even I didn't know what would happen, didn't know that I would end up where I am right now. But I am happy that I got the opportunity to work with such an amazing team."

Learn more about how Ruchika and the rest of Amii’s Industry team are helping organizations use machine learning to solve their problems. Visit our Industry Solutions page.