Photo by the Author
The obvious Getting started
Are we all in a race to the bottom of our own making? Data scientists have been employed for years to develop large scale language models (LLMS).
Now, the number of open data positions seems to be decreasing every day. Of those advertised, most of them seem abysmal.
By Abysmal, I don't mean extremely low salaries or unreasonable technical expectations from candidates. No, I mean those vague shrips: “Working well with AI productivity tools,” “Able to ship high amounts of code,” or “Plus fast dynamic capabilities.” Translation: Chatbot is your best coding partner, no guidance, no standards, just coding.
Chatbot, our very own creation, is now reduced to nothing more than a sloppy copycat. It doesn't sound like a very meaningful or fulfilling job.
In this area, is it possible to find meaningful work?
The obvious What are vibe codes?
Andrej Karpathya Open it The inventor, coined the term “vibe codes.” It means you don't code at all.
What you do: You drink your Matcha Latte, vibrate, give orders to the Coding Chatbot, and copy-paste its code into your code editor.
What a chatbot does: IT codes, checks for errors, and debugs the code.
What you don't do: you don't code, you don't debug, and you don't show code.
How do they feel about such work? Like full-time brain rot.
What did you expect? You have provided all the interesting, creative, and problem-solving aspects of your work in Chatbot.
The obvious Vibe codes have been removed
“It's not too bad for weekend projects on a lost weekend, but it's still really funny,” Andrerej Karpathy said about the vibe codes.
Despite that, companies you can trust – those who don't think of their products as “weekend-friendly” – have decided that it's still a good idea to start coding the vibe.
AI coding tools are coming in, and data scientists are being thrown out. For those who stay, their main job is to talk and chat.
Work is done faster than ever. You meet limits that were impossible before. The power of pretending to be productive has reached a whole new level.
The result? Completed prototypes. Code that breaks in production. Data professionals who don't know why the code isn't working. Hell, they don't even know why the code IS to work.
Prediction: Experts who know for sure that code will come back into fashion soon. After all, someone has to rewrite that code written “on the fly” with a chatbot. Talk about efficiency. Well, you don't get much better than that.
But how do you survive until then?
The obvious How do you find meaningful work now?
The goal is very simple: do what a chatbot can't do. Here's a comparison between what AI can do and what you can do.
Of course, doing everything requires some skills.
The obvious Skills needed
Finding meaningful work in the age of vibe requires these skills.
// 1. Technical writing
Most of the applications you will encounter come with incomplete and confusing information. If you can convert that information into a technical specification, you will be informed to avoid conflicting assumptions and expectations in the development work. Technical specifications help align all parties involved in the project.
Here's what this skill entails.
Resources:
// 2. Understanding data flow
Programs don't just fail because of bad code. Arguably, they often fail due to incorrect assumptions about data.
It doesn't matter what the vibe is, one still needs to understand how the data was generated, transformed and consumed.
Resources:
// 3. Repair Production
LLMS cannot correct the error in production. That's where you come in, with your knowledge of logs and metrics to find the root causes of production incidents.
Resources:
// 4. Architectural thinking
Without understanding their architecture, the systems will be designed to work in production (fingers crossed!), but they tend to fail under real traffic.
Architecture considerations determine system reliability, latency, turnaround, and operational complexity.
Resources:
// 5. Schema & Design Design
Poorly designed Schemas and definitions of the systems through which they communicate can cause a domino effect: a chronic failure that leads to excessive migration, which leads to communication and convergence between groups.
Create a beautiful design, and create durability and prevent leakage.
Resources:
// 6. Awareness of performance
Programs always behave differently in production environments than in development.
Since the whole idea is for the system to work, you have to understand how the elements reveal, how failures occur, and what and where there are and where there are bombs. With that knowledge, the transition between development and production will be very painful.
Resources:
// 7. Necessary negotiation
“Prevention is better than cure” applies here, too. You can expect almost endless releases and rewrites if the requirements weren't initially defined well. It's hell trying to fix it when the program is produced.
To prevent this, you must skillfully intervene in the early stages of development to correct the dimensions, relieve technical constraints, and translate the clear requirements into technical terms.
Resources:
// 8. Review of code of conduct
You should be able to read the code not only its functionality but more broadly its programmatic impact.
That way, you'll be able to identify vulnerabilities that aren't visible in the build or test, especially in teams generated by AI, and prevent hidden bugs that would disrupt your product.
Resources:
// 9. Cost & Performance Judgment
Your job has financial and operational consequences. It will be very important if you show that you understand them by looking at the computer usage, latency, traffic, and infrastructure costs in your work.
That's more important to companies than building expensive systems that don't work.
Resources:
The obvious Real jobs that still felt meaningful
Finally, let's talk about real jobs that involve using at least some of the skills we discussed earlier. The focus may be shifted away from the coding itself, but some aspects of those activities may still feel meaningful.
// 1. Data science (the real kind, not textbook-only)
AI can generate code, but data scientists provide structure, consultation, and domain understanding to abstract and, often, ill-defined problems.
// 2. Machine Learning Engineer
AI can train a model, but what about data preparation, training pipelines, infrastructure maintenance, monitoring, failure handling, etc.? That's the job of a machine learning engineer.
// 3. Analysis Engineer
AI can write SQL queries, but analytics engineers are the ones who ensure accuracy and long-term stability.
// 4. Data engineer
Data engineers are in charge of data reliability and availability. AI can transform data, but it cannot manage system behavior, peak changes, or long-term data reliability.
// 5. Machine learning ops / data ops developer
These roles ensure pipelines run reliably and models remain accurate.
You can use AI to suggest fixes, but operations, system interactions, and manufacturing failures still require human oversight.
// 6
AI cannot really come up with anything new, especially not new modeling methods and algorithms; It can simply change what is already there.
Anything else, technical knowledge is required.
// 7. Data Product Manager
This job description should define what data or machine products are to be created, including translating business needs into clear technical requirements and aligning stakeholder priorities.
You can't hire AI to discuss scope or risk assessment.
// 8. Management, compliance, and data quality roles
AI cannot ensure that data practices meet legal, ethical, and trust standards. One needs to define the rules and work for them, that is the management, compliance, and data quality roles.
// 9. Data visualization / decision science roles
Data needs to be linked to decisions to be meaningful. AI can generate charts all you want, but it doesn't know what things are making a decision.
// 10. Top Data Roles (Principal, Staff, Lead)
AI is a great helper, but a bad leader. Specifically, it will not lead.
Decision making? Cross-Domain Leadership? Technical guidance? Only humans can do that.
The obvious Lasting
Finding meaningful work in the age of the vibe is not easy. However, coding is not the only thing that data professionals do. Try looking at job ads, even if they require vibe codes, they also require some of those skills that ai can still replace.
Nate receipt He is a data scientist and product strategist. He is also a self-proclaimed educator, and the founder of Stratascratch, a platform that helps data scientists prepare for their interviews with real interview questions from top companies. Nate writes on the latest trends in the job market, gives interactive advice, shares data science projects, and covers all things SQL.
