Life science career advice- 8 mins
Life sciences students often reach out to ask me for career advice. I’m always happy to help however I can (feel free to email me!), but over the years I’ve found that much of my advice is generic enough to relay publicly once, rather than privately N times. This post is my attempt to capture the union of many career advice emails.
I’m a PhD student. Should I do a postdoctoral fellowship?
Short answer: It depends.
A postdoctoral fellowship is a unique type of training position that doesn’t really exist in most professions. For the vast majority of entry-level jobs, you’re expected to have the skills necessary to do the job or to acquire them within a few months of starting. In the natural sciences, many skill sets are complex and take years to master, so it’s unrealistic to hire an entry-level employees for a role that requires skills distinct from their past experience.
A postdoctoral position allows you the opportunity to change fields dramatically and learn a new skill set, even if you won’t be proficient for 1+ years after joining. Even better, you get to choose exactly what you work on during that time and have a large degree of autonomy. That’s a pretty remarkable opportunity! Commensorate with the training benefit, a postdoctoral position is compensated at a much lower rate than a comparable industry role for a proficient contributor.
We can then view a postdoctoral position as a role that’s partially compensated in training and autonomy.s Given this framing, I recommend candidates pursue a postdoctoral role if and only if:
- You are strongly considering an academic faculty career. A postdoc is requried for this route.
- You are hoping to learn a new skill set that is very distinct from your current skill set.
If neither of those conditions are true, I recommend candidates try to find a role in the relevant industry setting rather than pursuing a postdoctoral fellowship “by default.”
Industry postdoc programs tend to be an interpolation between a permanent industry position and an academic postdoc – you have slightly less autonomy, and the fit for position also needs to be commensarely higher than in academic, but you’re comepnsated at a higher level and have access to industry level resources. I think the same general decision criterion applies.
I am an experimentalist, but I’d like to be a computational biologist. Where should I start?
Congratulations on your new interest! Selecting among many possible career paths is a non-trivial part of the struggle, and you’ve now completed that much. Now, get ready for a long journey.
Computational biology is an odd, chimeric disipline that spans the intersection of biology, computer science, and statistics. There is no one “right” way to train. You will almost certainly develop a narrow sub-specialty within the field over time. Even the seemingly niche disiplines of functional genomics, genome assembly, and genetic variation analysis are fields unto themselves, to say nothing of quantitative imaging, proteomics, metabolics, and many other high-dimensional data modalities. You won’t be able to learn how to do “computational biology” broadly. Rather, you’ll learn how to practice one of these disiplines and develop a set of skills that are transferable across domains, making it easier for you to learn a new disipline in the future.
Learn to program
The first thing any aspiring computational biologist should do is learn to program. There’s no way around this – you absolutely need to be a strong programmer to be a good computational biologist. Building software is how you turn ideas into tools or run discovery experiments in computational biology. It’s the medium through which all your other skills must pass.
I recommend experimentalists start by learning Python, a general purpose language that underpins much of the modern computational biology ecosystem. Some courses will start you using R or Matlab. That’s okay, but these are more akin to “mathematical computing environments” than normal programming languages, and I’ve seen students misunderstand key computer science concepts based on the quirks of these ecosystems.
To help accelerate your learning, I recommend you set an a well-defined goal that’s associated with your current research project. This should be something that is definitely possible like “automatically make plots from my flow cytometry FCS files”, rather than something that includes any sort of scientific risk like “train a machine learning model to predict disease states from accelerometer data.” As you execute on the project, you’ll naturally encounter gaps in the programming knowledge you developed through a more formal pedagogy and work to fill them. After you’ve finished, you’ll not only have learned, but you now have a concrete software product you can showcase to future colleagues to demonstrate your programmatic skills.
How should I get my first job in industry?
What’s it like to work in biotech? Is it worse than being a PhD student/postdoc in academia?
Can I have a job at your company? (or, How to Network Effectively via Cold-Email)
I get a fair number of emails that ask fairly explicitly for help landing a position at my employer without offering much in the way of motivation. I try to use these as teachable moments for the author. I think there’s a fair amount of advice out there suggesting that you “get a job by networking,” that fails to explain what “networking” actually means or how it adds value. This misunderstanding is perhaps even further amplified in scientific careers by the fact that the academic job market is idiosyncratic. Your first job (Ph.D. position) is decided by a committee, unmatched to particular projects you will work on, and your subsequent postdoctoral recruitment likely began with a cold email, unlike recruitment in most other industries.
Networking for scientific roles is actually quite a useful process for both hiring managers and job seekers when done correctly!
I’ll digress for a moment to describe the talent market. Most products we buy and sell are fungible. One banana bunch at the supermarket is (roughly) as good as the next. There’s a clear legible pricing system and the buyer is clear about what they need (yummy bananas) and what they’re getting (bananas in various states of ripeness). It feels icky to talk about talent as a marketplace, but economically speaking we buy and sell labor in a manner that’s not so different from how we buy and sell bananas.
The key difference is that talent is not fungible! No two candidates are alike – each has unique skills, a unique training background, and unique communications style. When you look to hire a candidate to fill an open role, it’s both difficult to define exactly what you’re looking for (beyond “someone who will solve me current problem”) and difficult for the candidate to describe exactly how they can help your business. Both your needs as a hiring manager and the capabilities of the candidate are illegible by default. It takes a great deal of effort for each of you to determine if the role and candidate are a good fit for one another.
Cold-emailing a hiring manager is a form of networking, albeit a loose one. The purpose of networking as a candidate is to make your capabilities legible to hiring managers. What exactly does that mean? It’s almost impossible to convey the full-range of skills you possess and the full range of problems you can solve in a 2-page CV. You might have a bullet point like “Completed important project X at company Y,” but if I don’t realize that “project X” maps perfectly to my company’s problems, I might miss out on recruiting you for the role! Your goal as a networking candidate is to (1) understand what the company’s problems are and (2) help employees – ideally, hiring managers – at that company understand how you can solve their problem.
It’s very unlikely that you’ll find a good match for an industry role by cold-emailing a manager without offering any further motivation or details. At any time, a well-known company has hundreds of CVs from job-seekers. Simply sending your CV to a hiring manager’s personal email doesn’t differentiate your application from the others.
Rather, if you do decide to send a cold email, you can differentiate yourself by:
- Demonstrating that you’ve made some effort to understand the company’s problems
- Explaining how your past work prepares you to solve them
- Inquiring about any unknowns that might help us assess if we’re a mutual fit for a role
It sounds simple – and it is! – but I promise your emails will be infinitely more effective if you provide this sort of rationale and motivation, rather than simply describing your career stage, attaching your CV, and asking if there are suitable roles.