I saw a Reddit post recently about a biostatistician comparing pharma and tech compensation.

It was not a dramatic post. The person had an offer from tech, another from pharma, and sat down to compare the whole package instead of only looking at the biggest number: base salary, bonus, stock, pension, stability, work-life balance, and interest in the work.

In the end, pharma made more sense for them.

The comments were more interesting than the post itself. Some people stayed in hospitals because they liked the work and could see the impact on patients. Some stayed in academia because the team was good, the hours were flexible, the commute was easy, and their life already worked. Some wanted the higher-paying tech role but could not even get past the first screening.

The exact numbers were not what stayed with me.

What stayed with me was the question underneath all of it:

What kind of life does this work actually give me?

I have been asking myself something similar about bioinformatics. Not exactly pharma versus tech. Not academia versus industry. More like:

Do I still want to be a bioinformatics person?

And the answer, if I am honest, is probably no. Or at least not in the way I used to imagine.

This is not a dramatic goodbye. I do not hate bioinformatics. That would be too easy.

The truth is quieter than that.

Back in 2019, when I first looked at bioinformatics from Vietnam, the field felt almost imaginary.

There was interest. There were students. There were labs. There were people who knew biology was becoming more computational. But as an actual career path, it barely existed.

There was no company I could point to and say:

This is where people like us work.

Maybe that was why leaving felt necessary at the time. If the field did not really exist around me, then maybe I had to go somewhere else to find it.

So I went deeper into academic bioinformatics.

For a while, I tried very hard to make that path work. The science was real. The effort was real. But the life around it was not something I wanted to keep living.

I worked too much. I kept trying to prove something. Somewhere along the way, I started feeling that what I was proving did not matter very much outside a small circle of specialists.

That was a hard thing to admit.

When you are inside academia, it can feel like leaving means failing. It can feel like you are giving up before the work becomes meaningful.

But sometimes leaving is not failure.

Sometimes it is the first honest decision you have made in years.

I left academic bioinformatics in 2023. I left Korea and came back to Vietnam, not with a grand plan, but mostly with the feeling that I needed to start over.

Vietnam looked different from how it did in 2019. The field was still small, but it was no longer imaginary.

There were more companies than there used to be. Not many. Depending on how strictly you define bioinformatics, there are probably only five or six that most people in the field would recognize.

That may not sound small until you spend a few years in it.

The community is small enough that I can name nearly all of them, and small enough that I know people working at several of them. The same names keep appearing. Someone leaves one company and joins another. Someone becomes a hiring manager. Someone starts consulting. Someone moves overseas. Someone leaves biology entirely.

After a while, it starts feeling less like an industry and more like a small community.

That changes how a career feels.

One thing I did not understand back then was how strange bioinformatics is as a profession.

From the outside, it looks global. The papers are global. The tools are global. The conferences are global. The problems are global.

But the work is often local.

The knowledge travels easily.

The data does not.

Clinical data is sensitive. Human genomics data is sensitive. Hospital data is sensitive. National biobank data is sensitive. The regulations, contracts, and institutions around them are local.

And if the data does not move, the jobs do not move as freely either.

You can spend years learning global tools and global skills, only to discover that your opportunities are still shaped by a surprisingly small local ecosystem.

Global in knowledge.

Local in opportunity.

When you are trying to build a life inside that contradiction, those abstract words become practical. They become rent, salary, whether your skills can move, and whether you can grow without leaving the country, leaving the field, or accepting a role that only uses half of what you know.

One of the unexpected things that happened after I came back was teaching.

I taught students who had just entered university. I taught doctors. I taught people from completely different backgrounds. Some sold medical devices. Some sold chemicals. Some worked in hospitals. Some were trying to move into data work or tech.

The details were different, but the pattern felt familiar.

They all wanted something. Maybe better pay. Maybe more stability. Maybe a future that felt larger than the one directly in front of them.

A lot of them saw bioinformatics as a bridge: biology on one side, technology on the other. If they could just learn enough programming, enough Linux, enough statistics, maybe they could cross it.

I understood that feeling because I had felt it myself.

What I was less sure about was whether I should encourage it.

Not because bioinformatics is useless. The field is important. But I had started seeing how difficult the path actually was.

I remember students trying to learn Python after a full day in the lab. Students struggling through English documentation because most of the useful material was not available in Vietnamese. Students fighting package installation errors before they had even written their first useful script.

People who were smart enough, but had been given the wrong ladder.

That experience changed how I think about career advice.

For years, the message was simple:

Learn bioinformatics.

Biology is becoming data.

This is the future.

The first part turned out to be true. Biology did become more computational. The data kept growing. The tools kept multiplying. The scientific need never disappeared.

What I am less certain about now is the second part.

A growing scientific need does not automatically create a healthy career path.

Data still needs money. It needs infrastructure, companies, long-term ownership, and people who maintain the boring parts after the exciting announcement is over. It also needs enough opportunity for the people doing the work.

If a student asked me today whether they should enter bioinformatics, I would not give the same answer I might have given years ago.

I would not say:

Yes, just learn Python and R.

I would ask where they want to live, whether they are willing to move, whether they genuinely enjoy biology, whether they genuinely enjoy software, and whether they are prepared to compete in a field that is global in tools but local in opportunity.

I would also ask whether they understand how much the field may ask from them.

Because bioinformatics has a habit of quietly expanding. You start by learning biology, then statistics, programming, Linux, workflow systems, cloud platforms, data management, debugging, and eventually all the things nobody mentioned when you started.

Maybe that is why the field is often misunderstood.

People only see the part closest to them: the biology, the statistics, the software, or the infrastructure.

Very few people see the whole thing.

Later, I ended up working on the platform side of the industry. Not doing biological interpretation. Not publishing papers. Supporting the systems that researchers relied on.

That experience changed my view of the field more than anything else.

Researchers would arrive frustrated because the platform was difficult. Many were new to cloud computing. Many had never managed infrastructure. Some principal investigators delegated everything to students and were shocked when costs exceeded expectations. Some workflows failed after days of runtime. Some analyses worked until a dependency changed. Some environments became impossible to reproduce.

Behind every scientific question was a mountain of engineering nobody wanted to think about.

Modern bioinformatics depends on an enormous amount of hidden machinery: storage systems, permissions, workflow engines, containers, compliance controls, audit trails, cloud infrastructure, and data governance.

Most researchers never wanted to become experts in those things.

But the science increasingly depends on them.

That is where I started to understand why some difficult platforms still survive.

Not because they are pleasant. Not because they are easy. Sometimes they survive because they satisfy legal, security, and governance requirements that ordinary tools cannot satisfy. Sometimes a platform exists not because researchers love it, but because the data is too sensitive for simpler arrangements.

That creates a strange kind of frustration.

The platform is hard to use. The users are frustrated. The support team is frustrated. The researchers only want to do science. The students are suddenly expected to understand cloud computing, billing, storage, permissions, and workflow execution.

And yet the platform exists because the alternative may not be legally or institutionally possible.

That is bioinformatics at scale.

Not clean. Not elegant. Not what the tutorial promised.

But real.

The same is true of the tools.

The field runs on imperfect solutions. Workflow engines can feel patched together. Pipelines can depend on a few people knowing how everything works. Scripts can grow from Bash, Python, Makefiles, YAML, containers, and hope.

But those imperfect tools are often still better than every lab inventing its own private mess from scratch.

That is one of the strange things about bioinformatics.

It keeps moving through compromise.

Researchers keep publishing. Students keep learning. Engineers keep patching. The next workflow arrives, then the next dataset, then the next generation hoping things will be better.

And somehow the field survives.

For a long time, I thought the combination of everything was what attracted me to bioinformatics: biology, statistics, programming, data, tools, and systems.

Eventually I realized I was becoming interested in a very specific part of it.

Not the biological interpretation.

Not the scientific conclusions.

The machinery underneath.

The workflows. The platforms. The automation. The reliability. The operational details.

A failed analysis can look like a science problem from the outside.

From the inside, it is often a systems problem.

Those were the problems that kept my attention.

When a workflow failed, I was less interested in the result than in understanding why the system allowed it to fail. When users struggled, I was less interested in writing another page of documentation than in making the failure harder to trigger in the first place.

At some point, I stopped asking only whether the analysis was correct.

I started asking different questions.

Who owns this workflow? Who maintains it? Who can reproduce it? Who will still understand it six months later? Who pays when it fails after three days? Who explains the cost to the lab? Who fixes it when it breaks again?

Those are not biology questions.

They are systems questions.

And I think that is where my attention has been moving for a long time.

This is not a post about bioinformatics being dead.

It is not.

The science matters. The work matters. The people certainly matter.

I do not think everyone should leave. Some people are exactly where they should be, and I respect that.

But I no longer think I am one of them.

I do not think I am leaving because I failed at bioinformatics.

I think I followed it far enough to discover which part I actually care about.

Bioinformatics gave me a career. It gave me difficult problems. It taught me what real scientific computing looks like, how messy data is, how fragile research software can be, and how much invisible engineering sits between a scientific question and a usable result.

And somewhere along the way, it taught me something else.

I came into bioinformatics because I liked science and code.

I am moving on because I found out I like systems more.