How might you use machine learning in a medical career?

If you’re a healthcare professional with an interest in machine learning (ML), but unsure how can combine it with your career, then we hope you’ll find this post useful.

The most common ways of combining a medical background with ML is in academic research (both novel uses of ML in healthcare as well as augmenting existing research with novel ML techniques) and in industry (by providing clinical input to health technology start-ups and companies). It’s worth noting, also, that there’s a fair overlap between the two.

Less common roles include consulting, to advise companies on their use of ML, starting your own start-up or company, or becoming a research engineer / machine learning engineer (a much more technical role).

In this article, we’ll outline the roles, the rough requirements and potential next steps for those considering a particular path.

For suggestions on what and how to learn about machine learning, check out this article.

Machine Learning Healthcare Research

This is one of the most exciting and active areas of medical research at present. There are many research groups working in the area, and many more continuing to pop up. The groups usually consist of collaboration between clinicians and data scientists / research engineers. There is significant value that can be added by a clinician who understands the relevant machine learning aspects for projects.

Rough Requirements:

  • Medicine: broad understanding (i.e. year 3 med school+ )
  • Machine learning: fundamentals + area specific to research
  • Other: research experience helpful (not essential)
  • Read key papers on Pubmed and Arxiv
  • Join a project or research group - see our post for suggestions how
  • Set up a research group (if more senior clinically)

Clinical input to a start-up or company

There are an increasingly large number of companies working to use machine learning in healthcare in one shape or form. For example, the pharmaceutical industry is looking for more clinical and machine learning expertise. Clinicians who can “speak their language” (of machine learning) can add a lot of value by translating their clinical insights into the language of data. The types of input can be vary varied; from product design to user experience to safety and regulations. Roles can be full-time or part-time, meaning that it can be maintained alongside clinical practice.

Rough Requirements:

  • Medicine: depends on company, for many minimum ~1 year clinical experience
  • Machine learning: breadth of my course and depth of Andrew Ng courses
  • Other: business experience helpful

Consulting

Consulting has become a popular “exit-route” for doctors and healthcare professionals, with disillusionment a common factor that people cite for the career path change. There are several different types of consulting, with ‘strategy consulting’ being one of the most common. As someone with experience in healthcare and an understanding of machine learning (and technology more broadly), you can find a role advising other companies on how to integrate the two.

Rough requirements

Depends on level (graduate vs experienced hire) and type of consulting (strategy vs operations vs other) and industry (healthcare, finance, etc)

  • Medicine: For healthcare consulting, experience working in a healthcare system. For others, non-essential (emphasis is on soft skills)
  • Machine learning: If consulting in technological area, the fundamentals, and understanding higher-level impacts (e.g. real-world case studies)
  • Other: business experience and understanding is useful

Starting own start-up, company or initiative

This is a much less-defined role than the others discussed here. Start-ups can take many shapes or forms. There is not a well-defined path into starting your own thing, although experience in the relevant areas are important. Both machine learning and healthcare are fairly technical areas, so often require several years working in the area to gain the necessary insights. Fundamentally, you need to understand a problem and how to go about trying to solve it.

Rough requirements:

  • Medicine: No hard requirement – as much as required for the insight
  • Machine learning: Same (no hard requirement – as much as required for the insight)

Ideally 2+ years working in role exposing you to both ML and how it could solve clinical problems

No well-defined path. Gain experience, think of ideas and go for it. Iterate throughout.

Research engineer / machine learning engineer

These positions involves actually coding and implementing new machine learning tools, which can be in healthcare or otherwise. It’s very rare for someone to make the transition from medicine to these roles as it requires a high level of programming expertise and a deep understanding of machine learning.

Rough Requirements:

  • Medicine: medical degree perhaps useful, but full degree not required
  • Machine learning: Ideally Bachelor’s/Master’s in Computer Science/Machine Learning, plus practical experience of coding solutions for real-world problems
  • Take a degree (typically Bachelor’s/Master’s in Computer Science/Machine Learning)
  • Code mini-projects
  • Work in a junior role (e.g. junior data scientist)

We hope this is helpful!

If there are any options you think we’ve missed, or any other suggestions, please comment below or email us at career@chrislovejoy.me.