I got hired a couple of months ago by a product-based company.

All I've done these two months is learning how the product is made (it's a monolith), how everything is coupled together and how to integrate some feature they asked me to do.

I developed this feature itself in less than a week, but the other two months have been basically doing reverse engineering by spamming debug on this monolith.

I come from a consultancy firm in which the work is more bullshit-based, smoke and mirrors, etc., but I worked almost everyday with data science tools (huggingface, pandas, keras, building my own models, dealing with my data pipelines, model and data versioning, plots, etc.), in short, more what I'm supposed to do.

Is this normal? Am I just overreacting and I should have more patience? Should I mention it to my manager?

  • 4
    If you have been reverse engineering their product for two months, has that given you any insight into ways in which the data it throws off could be more fully exploited? A data scientist looking at virtually any enterprise system and product workflow should potentially be able to spot fallow data resources that could be harvested with the right data science projects. Propose those projects.
    – tbrookside
    Commented Aug 24, 2022 at 12:00
  • Was the developing of the new feature more like data scientist work or more software development? If it is the first, then what you are currently doing is just the training to be able to do that actual work later on.
    – quarague
    Commented Aug 25, 2022 at 9:55
  • We could say this feature is slightly related to data science, but still, the time dedicated to actual data science in this task has been less than 1%. I'm concerned because even if I know that actual data science part in real-life projects with products is relatively small (proceedings.neurips.cc/paper/2015/file/…), but this is too much, to the point that I feel that my data science skills can be impacted Commented Aug 25, 2022 at 10:24

3 Answers 3


Is this normal? Am I just overreacting and I should have more patience? Should I mention it to my manager?

Whether its normal depends on the company. In some companies, a new employee can dive right in fairly quickly into their role. In other companies it can take months to begin to be assigned work directly related to your role. The extreme case is that you never end up doing the actual work that you were hired for.

If the product is a monolith as you have described, then it may be normal that your company is taking their time before they assign you the tasks directly related to your role, as they may feel you need a good understanding of the product first.

Regardless, you should speak to your manager and ask what the roadmap is for your transition into work that is actually a part of the role you were hired for. That should give you some insight as to where your career is heading within this company.


This is a clear case of you needing to set up a meeting with your manager to talk about expectations. Ideally, you should have talked about things like this during your onboarding and first few weeks/months. Just talk about your expectations for the role and how they are not being met, your managers expectations for you and draw your conclusions.

It can be that your manager just expects you to propose/start working on new projects, or that the manager is just happy somebody is doing the grunt work. We don't know.

You are a senior data scientist, so you at least should be able to answer the 'is it normal' question. But that is not the important question. If your previous job made you happier, then you should reconsider the current one if the talk with the manager doesn't end up with confidence on your side that things will become better for you.


As others have said, speak to your manager about what you're interested in in terms of development Vs what the company needs and see if there is overlap.

I wanted to add some colour based on being in the "data industry". The title senior data scientist is a very vague one. In many companies you are not involved in machine learning for much of your time but work more as a data analyst. The role is to figure out what features are working and what isn't. This involves understanding experimentation (A/B testing) as well as statistical significance. It also involves running regression analysis in cases where experimentation is impossible (e.g. TV advertising). This is what makes it a "scientist" role. You can do all of this by using SQL and never touching a ML framework. In some companies you can spend time building a machine learning model but this is often shallow e.g. doing some clustering and writing a report about it. Actual machine learning roles are more often called ML researcher or ML engineer. These tend to be bigger tech companies that want people with computer science degrees / PhDs.

The reality is that it doesn't make sense for most companies especially startups to invest heavily in machine learning. ML takes fixed cost to build but delivers % improvements in value. So it makes more sense for Google than a small tech company. Consultancies can actually get more of this as they get hired by big stupid companies to do "cool stuff" that they probably don't need and could fix by effective engineering.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .