Moving to the Dark Side

Leaving the Pipette for a Keyboard.

Part II - Lost in translation

In this section I will try to translate, or at least give enough clues on how to read a job ad from an company. The approach I prefer is to pick one or two real job postings and take them apart line-by-line (sort of). I will anonymize the postings as much as possible because I want to keep the analysis fairly generalizable. A lot of this post is guesswork because I didn’t write any of these nor are these from my employer, so take it all with a pinch of salt.

Before I pick my metaphorical scalpel, let me re-assure of two things when you take a job in the industry:

  • You will do what you were already doing in academia, by and large. Depending on the role there will be a special focus on some tasks, in detriment of others, but your day to day will probably not be vastly different (with some exceptions1).
  • If there is X skill you don’t have to get the job done, don’t worry (be happy2). You have the most important skill which is the ability to learn and to deal with setbacks. That’s what academia trained you for.

Example 1 - You will become a scientist

Scientist* Bioinformatics

About the job Become a member of the XXXX Family!


Let’s start with this. What is a “Scientist”? Weren’t you one in academia?3

Titles (and levels) will vary depending on the company but generally it will be roughly along these lines:

  • Research associate, someone with a BSc/MSc, equivalent to a lab technician4
  • (Senior) Scientist, PhD holder. Experience will dictate whether one be hired at senior level or not. Fresh out of the PhD is likely to be “just” Scientist, but looking in the posting for years of experience will give you a good idea if you are “senior”.
  • Group leader / Lead / Principal Scientist. Here it starts to get murkier with titles. But it usually denotes someone very independent, with not just a good technical level but also a good understanding of what needs to be done to advance the business. It is also at this level and higher that generally people start becoming either individual contributors (lead technically) or managers (lead a team).
  • Director / VP etc. Usually require years of industry experience, so unlikely you will get one right out of academia after a PhD or a PostDoc(exceptions apply to former group leaders, etc)

I can’t talk about all the jobs titles, or how each company deals with levels, but looking at the experience requested gives you a pretty good idea of what level is expected.

But did you notice that sneaky “*” in the job title? I am not sure what it means exactly, but there is often some flexibility in the hiring (contract). Meaning that depending on your experience you might become either a Scientist or a Senior Scientist, which the asterisk might be denoting. Importantly, and spoiler alert, this job does not ask for a minimum number of years experience in the Industry which means it’s a great entry-level position.

The Company

As a part of our team, you will play a key role in developing solutions for some of the most crucial scientific challenges of our age. We develop treatments following the highest scientific and ethical standards – writing medical history.

We aim to reduce the suffering of people with life-changing therapies by harnessing the potential of the immune system to develop novel therapies against cancer and infectious diseases. While doing so, we are guided by our three company values: united, innovative, passionate. Get in touch with us if you are looking to be a part of creating hope for a healthy future in many people’s lives.

This is all a bit corporate, but look at it as the company showing what they can offer in high level terms. What does the company takes pride in? Does it match your interests?

You will also need to learn about the company for both the application - don’t send a template cover letter, tailor it to each job application - but also for the job interview. If the applicant didn’t bothered to read and understand the intro in the company’s website the interviewer will pick it up very quickly, and it will not be a positive note.

Read between the lines

This sentence is not very prominent and yet could tell a lot about the role:

At our company you will support the Bioinformatics R&D team with the ongoing bioinformatics research.

This tells you that this company has a group doing research and development (R&D) specifically in bioinformatics. This probably means less of a support role, like an internal Core Facility, but rather developing tools / pipelines that solve problems across teams.


Your main responsibilities are:

  • Architecture and implementation of software solutions in Python including validation (unittests)
  • Concept development for algorithms in the field of Artificial Intelligence, use of AI models in production and maintenance of AI models
  • Development of software and pipelines for processing and analysis of NGS and other high-throughput data
  • Contribute to innovative projects with different partners such as biological and machine-learning teams to collaboratively develop new analytical tools
  • Depending on your experience you will supervise students as well as colleagues
  • Literature research on innovative methods in the field of NGS, AI/ML and bioinformatics

This broadly what you will do on a daily basis. In sum, your job!

As you can see, the tasks seem to revolve write good code (in python), and do a lot of method development in ML/AI. There is also the bit about reading papers which makes it even more like an Postdoc except that it will be in a team and with a clear goal.

To me it reads like a position where someone will write a lot of code and research new algorithms. What I imagine the day to day to be like is a lot of discussion with many teams to find out where data analysis can be improved, or how the company can extract more knowledge of existing data. Or maybe there is a key biological problem that experiments alone won’t solve and data science comes to the rescue. After developing a project / strategy the hire will start investigating how to find a solution and putting into practice. It is also possible that there already a few ongoing / planned internal projects and you will start working on those. There will probably be cycles of research followed by coding, evaluate the results and either develop a tool / pipeline or go back to the beginning.

In sum, this reads more like a research job than a service facility. Regardless, any doubts you might have about the role, the hiring manager will probably explain it in your first interview session.

Skills to pay the bills

What you have to offer:

  • PhD in the field of bioinformatics, mathematics, biotechnology, biology or a comparable discipline
  • Experience in computational biology, bioinformatics, algorithm development, AI
  • Experience especially with Python, but also knowledge of other programming languages such as R
  • Solid Experience in analyzing and interpreting high-throughput processes incl. NGS data sets would also be highly beneficial
  • First experience in the analysis and preparation of clinical datasets
  • Practical experience in a biological/medical field preferred. Ideally sound knowledge of immunological and molecular processes in cancer cases
  • High self-motivation and a solution-oriented approach as well as distinct ability to work in a team

This section is actually pretty clear (that’s not always the case). The list of skills is short and to the point - I’ve highlighted what I see as the key aspects but even in those there is some leeway (italics).

You probably heard that if you have 50% of the skills listed in a job posting you should apply and I agree. However, there is always one or two that are critical and without which it’s unlikely that an application will be successful. In this example, one might think python is one of them, but maybe it’s the algorithm development and if you show you are really good at that, but you did it in Rust, and you have a dusting of experience in python, your application might pass the first hurdle. If you can do it in Rust, you can probably learn python (that’s me putting myself in the mind of the hiring manager).

Which are the key skills is sometimes difficult to judge. What I did here was to have a guess, but it can be completely wrong and it’s very unlikely we will ever find out. Speaking of which…

Story time.

I once applied for a job that was about a 90% match with my skill-set. I had the experience, the proven mindset, knew the biology. I got interviewed, did well, but didn’t get an offer. Through a friend that worked in this place, I got to know that the candidate that was hired had less experience than me in most requirements, except in the one technology for which he was an expert and I was a novice. This was the main skill the hiring manager wanted! I never had a chance, though reading the job description one would have never guessed.

The conclusion is that you should apply if your experience / skills partially match the requirements, and are able do the job as described. Don’t take it personally if you don’t get an offer. There are many unknowns in hiring.

Example 2 - Read all of the papers

(or use a computer to do it for you)

This position is very unusual in my view, but it serves as a good example to show that there is a position that is just right for you. Even if your experience is somewhat left field. The right job might not be available now, but it’s out there. It also exemplifies one of many roles someone coming from academia can take in the industry, leveraging their skills but not necessarily doing exactly what one does daily in an academic position. And it’s a data science position with few of the classical data science skills required.

It’s a long one so I’ll try to break it apart into bite size chunks - large chunks, like a Chunky Ice Cream.

A what now?

Biomedical Literature / Patent Junior Data Scientist

The (..) group has a mission to unlock information that makes cures possible. This team builds tools and provides services that help our scientists interact with, manage, and consume knowledge from scientific literature and other sources to foster innovation in drug discovery and development. We leverage leading technologies and methods to provide scientists with our industry’s best capabilities around literature-based knowledge discovery. Our team has the scientific and technical competence to use a very large number of knowledge sources to provide high quality answers to our customers.

This Biomedical Literature / Patent Data Scientist role will be part of the XXX Research groups. The team supports Technology and Therapeutic Area teams in Discovery and Development Sciences in finding, (text and data) mining, and analyzing published knowledge from literature, patents and other sources and building innovative publication data science solutions. The Literature / Patent Data Scientist should have a sufficiently strong scientific and technical background and information resource knowledge to scope and fulfill these services. Additionally, the Literature / Patent Data Scientist will effectively collaborate across teams to deliver impactful services and innovative solutions to our internal research teams. The role will help build new analytical capabilities and value-add tools for target evaluation and lead compound discovery leveraging the variety of biomedical data from databases & knowledge graph as well as generating new data and insights based on publication mining.

This is a long way of saying they have a product that might be used internally but it’s also sold to customers. The product is knowledge extracted from papers.

But isn’t that what a scientist does all day, you ask. Yes, but (i) not with code (unless you work on NLP); and (ii) time to go through papers is expensive (at least in the industry). In pharma it is beneficial to get as soon as possible to an hypotheses (target) on how to cure a disease. If someone offers a product that speeds up the process this is very valuable.

Before we goo deeper in the job description, There are a couple of expressions here which I rarely heard before taking a job in the industry:

  • target evaluation, a target is a gene / protein which can is candidate to find drugs to cure a disease. Think BRAC for breast cancer or APP for Alzheimer. But just because it’s implicated in a disease doesn’t mean it’s a good target for drug development, or maybe the link to disease is circumstantial. Making sure it is a good target is essential to develop assays correctly (and save money).
  • lead compound, this a drug found in a screen that reverts the disease phenotype in an in vitro assay. It then goes on to be developed further until it becomes an approved treatment (maybe).

Once you understand the drug discovery process it becomes clear why using computational methods to automatically extract information (and knowledge) from the literature is extremely valuable - and we, bioinformaticians, can contribute a lot to it.

But where do Patents fit in all of this? Well, imagine a company has a great target, or idea for a diagnostic kit, but it has been patented by someone else, or there is an approved drug for another disease that acts on the target. Is it worth it to follow up on that target?

This is were business analysis comes in. Maybe the idea is not exactly the same as the one patented, or it could be modified just enough to avoid infringement. Or maybe it’s so great that it’s worth buying the patent. Regardless, this knowledge of an existing patent is very important for an informed business decision.

The job

Key Responsibilities

  • Leverage scientific domain and technical knowledge to support scientists with identifying most relevant information, published knowledge as well as decision-making in R&D projects and deliver customer specific search & analysis solutions. Key focus: Target evaluation (identification, validation, prioritization) and Lead discovery support.
  • Support role scientific communities in Knowledge Discovery Projects.
  • Fill role as supportive business partner for (Immuno-)Oncology teams.

I am guessing that in this role someone will be assigned a customer / project and the task is to use their biological knowledge to come up with disease targets or evidence to support those and / or for experimental design.

  • Use ontology-based text data mining, scientific data analytics, APIs, and other tools to drive end-user engagement, improved R&D efficiencies, and reveal new discovery opportunities across therapeutic focus areas.

And this how the above will be accomplished: query databases or use other tools, including scientific knowledge to give customers good leads. “R&D efficiencies” here stands for making drug discovery faster.

It also expected that new projects will be created, meaning, not only complete the assignment but also coming up with new targets or insights - sounds like what a scientist does, no?

  • Provide unique scientific insights and expertise by designing and developing solutions for finding, extracting, curating, and visualizing knowledge from scientific publications, with focus on literature and patent analysis, and combining this data with other relevant structured biomedical data.

This is where some coding ability and knowledge of biomedical (and patent databases) are handy.

  • Collaborate with internal data scientists, information scientists and software developers to design and implement tailored solutions to meet well-defined stakeholder requirements with focus on early research.

Teamwork and support. This collaboration with data scientists is specially good for someone who isn’t very proficient in some aspects of Data Science (yet) but might have the opportunity to learn through collaboration.

  • Monitor continuously newest technology trends relevant for publication analysis and knowledge discovery and establish an external presence through conferences and publications

Be on top of the literature and network at conferences. Sounds familiar?

Skills to pay the bills

  • Bachelor’s Degree or equivalent education with typically 5 years of experience, or Master’s Degree or equivalent education with typically two years of experience in Bioinformatics, Cheminformatics or life sciences (Biology, Pharmacy, Medicine, Pharmacology, Biochemistry, Medicinal Chemistry or related disciplines).

Straightforward. No PhD required but almost.

  • First biomedical knowledge and ability to translate complex scientific questions from research scientists into information solutions. Good basic scientific understanding of key focus areas (Oncology or Immunology) and interest to develop deeper knowledge desired.

A translator skillset - I love these. Basically you will need to be someone who knows a bit of biology and a bit of computer science to bridge between these two domains. And this is harder than many think because it’s not just about knowledge, but also communicating effectively.

  • First programming and informatics/data science/computer science background: Proficiency in Python required. Ability to understand and modify existing code as well as develop new scripts and set up new data processing workflows.
  • Experience using relational databases and SQL, basic understanding of knowledge graphs, Cypher language and graph databases (neo4j) would be an additional asset.

The top one (python) seems like a fairly hard requirement, though the level of proficiency might be open to interpretation. Decent python but good coding in another language, which indicates ability to learn programming languages, might be enough.

  • Experience with biomedical information resources (e.g. PubMed), literature and/or patent research, information analysis and data normalization is mandatory, advanced knowledge on patent analysis is a plus. Experience in text mining, natural language processing, semantic enrichment, ontologies, data mining or machine learning/AI is a plus.

Long paragraph but only one key skill! I am not sure if it’s a single skill or laundry list of many skills. knowing how to parse Pubmed and basic data analysis? Just apply if you have one of those.

  • Familiarity with early drug discovery process (target evaluation, pathway analysis or lead compound discovery) or prior experience in pharmaceutical early research is huge plus.

Knowing the drug discovery process is very important for pharma / biotech jobs and usually only comes with experience in the industry. A catch-22. However not a key requirement in this case.

  • Analytical skills to process, analyze, visualize, and present results. Knowledge of technologies for data analysis and visualization of complex data (e.g. Rest APIs, XML, JSON, GraphPad Prism) as well as knowledge of public domain standard biomedical terminologies (e.g. MeSH, HGNC) is desirable.

This is very long list of nice-to-haves. Fundamentally, knowledge of APIs and biomedical DBs, along with ontologies. For me it boils down to data analysis skills, ability to parse data, and use GUI tools for the analysis.

  • Systematic problem-solving, quick learner and superior attention to detail in developing tailored solutions, high degree of reliability and integrity.
  • Demonstrates an interest in working collaboratively, cross-functionally, and in inter-disciplinary teams with an ability to effectively communicate, both verbally and in writing to scientists and non-scientists.
  • Well-organized and balances taking direction from others with taking initiative to manage multiple projects and learning responsibilities.
  • Innate scientific curiosity, technical creativity, and innovative thinking, motivated to break new ground in the field of information/literature analysis and knowledge discovery.

Soft skills. To determine if you have some of these you will have to have a long, hard conversation with yourself - am I really organized? Others are just part of being in academia (scientific curiosity), but be prepared to give specific examples during an interview.

Final words

As mentioned, some of this is guess work, and understanding job posting sometimes feels like reading tea leaves, but I hope I give you enough insights on how to read them. If you have something to add, a correction, or have a sentence from an ad that you would like “translated”, feel free to comment below.


  1. I know people that retrained and are now investment advisors. This is obviously a big jump but some of the key skills learned in academia are useful such as communication and analytical ability. I sometimes feel like some of these skills are undervalued by candidates themselves. 

  2. I’ll be sprinkling “cultural” references a lot in these posts. Not sorry. 

  3. I once quipped that I had to leave academia to finally have the title “Scientist”. Oh the irony. 

  4. Unlike academia, in some companies, you don’t need a PhD to go from research associate to Scientist. It might take time, but experience is valued and rewarded. 

Newer >>