MFine is one of the leading telemedicine players in the country. As an on-demand, digital primary healthcare platform which offers professional diagnostics and health check-up services, the company partners with thousands of top hospitals, doctors, and diagnostics centres in India to bring high-quality primary healthcare services directly to consumers using artificial intelligence (AI) and mobile technology.
MFine’s data science team has been instrumental in everything it does today. “There are multiple projects that have evolved into high impact products,” said Ajit Narayanan, CTO at MFine. For example, he said their core consultation product that millions of its customers use is currently aided by its ‘virtual doctor,’ which coordinates more than 80 per cent of cases. Recently, the company has introduced a new SpO2 monitoring feature on mobile using the rear camera. Currently, in the beta stage, it can save many lives without any investment on the users’ part.
Sign up for your weekly dose of what's up in emerging technology.
In the past, the team has also delivered other solutions, such as the lab report reader, which extracts the lab parameters and the values from the lab report. It charts the users’ lab results longitudinally in the app. “There are many others within the app and ensure smooth consultation on the platform,” said Narayanan.
At MFine, the data science team is part of the product and technology team with about eight members. The company has subdivided the team based on the kinds of medical units it receives from its patients on the MFine platform. These happen to be computer vision (to look at images of patients’ problems or images of one’s reports), NLP (conversations between doctor and patient), and machine learning specialists (for reasoning).
Currently, the data science team at MFine operates in two models – insights and discovery. In the insights mode, the projects that are undertaken by the team are mostly based on the medical insights derived from the statistical analysis of millions of data points on the MFine platform. These projects not only contribute to the improvement of existing product lines but also enable the transformation of products to be driven by AI.
In the discovery mode, the team directly interacts with doctors and conducts exhaustive reviews of clinical literature to develop novel healthcare solutions that are futuristic. These projects are expected to evolve into business lines between six months to 2 years, depending on the market needs.
In an expansion mode, MFine is now looking to hire lead data scientists with four to seven years of work experience.
“The interview process varies according to the experience of the candidate we wish to hire,” said Narayan. For instance, entry-level candidates may be expected to demonstrate their hands-on programming skills. In contrast, more experienced candidates will be expected to present their previous work and explain alternative ways to solve the same problem in the first round.
This would be followed by a deep dive into the technology questions tailored to real data science problems they solve at MFine. Finally, the candidates are evaluated based on their brainstorming capabilities, approach to feedback, technical depth, communication skills and cultural fit.
At an entry-level, the data scientists should be capable of:
- Understanding and implementing state-of-the-art AI algorithms
- R&D of AI algorithms
- Developing proofs-of-concept and productizing algorithms
Nikhil Narayan, director of data science at MFine, said that the lead data scientists should have the responsibility to understand and translate clinical needs into a data science problem statement, architect AI solutions, gain interdisciplinary domain expertise, and interact closely with the DevOps and Product teams. Further, he said that the candidate must be familiar with the upcoming trends in MLOPs, mentor data scientists, publish patents and papers, handle clinical collaborations, and partner with industry and academia.
MFine assesses data science candidates based on ownership, expertise and impact.
“The candidates are responsible for proposing and owning their solutions from inception to completion and integration with products. Expertise relates to the domain knowledge that the candidates already have and pick up on the job. Nimble learning is one of the traits that we look for in candidates at the time of both interviewing and assessment. Impact refers to whether the solution that the candidate proposed directly contributes towards a business need or a research problem,” said Narayanan.
Dos & Don’ts
Narayan said that there is a lot of emphasis on mindset and behavioural traits at MFine. “We look for candidates who are not scared to fail, are capable of learning without inhibitions, can question the status quo, are helpful, are amenable to feedback, have an open mindset and do not work in silos,” he added.
Further, talking about the common mistake made by the candidate when interviewing for data science roles at MFine, Narayan said that many candidates assume that deep learning is the only way to solve a problem statement.
“Not many startups have a fixed model for work culture, but at MFine, we do have a good and proven model,” said Narayanan. He said their work culture follows an engineering model outlined by the Wharton School of the University of Pennsylvania. The company believes to have created an environment that fosters a performance-driven, achievement-oriented candidate and has interdisciplinary problem-solving capabilities.
Adding to this, Narayanan, who heads the data science team at MFine, said that their team has the freedom to pace their work according to their comfort as they believe a lot in the mental wellbeing of the team members. “Healthcare is perhaps the only domain where all modalities are used simultaneously to solve a task at hand. So, there is a lot of scope for multi-modal multi-task research and development work on some of the most challenging datasets one can ever encounter,” said Narayanan.
Further, he said that the problem statements are challenging and perfectly suit someone after gaining a large amount of knowledge in a very short span of time. “That is probably the biggest advantage of joining a startup where there are no boundaries between teams, and everyone can pitch in to solve a common goal,” said Narayanan.
Besides this, the company also offers other perks like free medical consultation for employees and family on the MFine platform, free food, flexible work hours, flexible work locations and good health insurance, to name a few.
What are you waiting for? Apply for data science and tech jobs here.