7 mistakes that scientists make, applying for a job

The data learning market is crowded. Employers and recruiters are sometimes real nature, which the spirit, when you thought you would start to negotiate your salary.

As if fighting competition, recruiters and employers are not enough, you also have to fight. Sometimes the lack of success in interviews really concerns data scientists. Making mistakes is acceptable. Not learning from them is nothing else!

Let’s review some typical mistakes and see how not to commit them when applying for a job in learning data.

1. Treating all the roles themselves

Mistake: Sending the same CV and cover letter to each role, from the position of ponderous tests and customers, to being a cook or Timothée chalamet.

Why does it hurt: Because you want this work, not the “best general candidate for all positions that we do not employ.” Companies want you to adapt to a specific job.

The role in the software startup may prioritize product analysis, while the insurance company employs for modeling in R.

NO Adjusting your own The CV and cover letter, to present itself as a very suitable position, has the risk of overlooking before the conversation.

Trouble:

• Read the description of the position carefully.
• Adjust your CV and cover letter to these work requirements – skills, tools and tasks.
• Do not exchange skills, but show your experience in the appropriate applications of these skills.

2. Too general data designs

Mistake: Sending a portfolio of data designs full of washed projects, such as Titanic, Iris Datasets, multiply or forecasting home prices.

Why does it hurt: Because the recruiters will fall asleep when they read your application. They saw the same portfolio thousands of times. They ignore you because this portfolio only shows a lack of business thinking and creativity.

Trouble:

• Work with sloppy, real data. Source of projects and data from websites such as StratascratchIN KaggleIN DatasfIN DataHub by NYC Open DataIN Amazing public data setse.t.c.
• Work on less popular projects
• Choose projects that show your passions and solve practical business problems, preferably those that your employer may have.
• Explain compromises and why your approach makes sense in a business context.

3. Undertaking SQL

Mistake: Without exercising SQL, because “it is easy compared to Python or machine learning.”

Why does it hurt: Because knowledge of Python and how to avoid excessive fit, you don’t make you an SQL expert. Oh yes, SQL is also strictly tested, especially for the roles of analysts and at medium levels. Interviews often focus more on SQL than Python.

Trouble:

• Exercise elaborate SQL concepts: thanks, CTE, window functions, connection ranks connection, rotary and recursive queries.
• Apply platforms like Stratascratch AND Code leetcode To practice questions about the real SQL interview.

4. Ignoring the product’s thinking

Mistake: Focusing on models instead of business value.

Why does it hurt: Because the model that provides for the client’s departure from 94% ROC-AUCBut mainly flags customers who no longer employ the product, they have no business value. You can’t stop customers who are gone. Your skills do not exist in a vacuum; Employers want you to employ these skills to provide value.

Trouble:

5. Ignoring MLOPS

Mistake: Focusing only on building a model, while ignoring its implementation, monitoring, tuning and the way it works in production.

Why does it hurt: Because you can stick your model, where you know-where you can’t employ it in production. Most employers do not consider you a solemn candidate if you do not know how your model will be implemented, retrained or monitored. You will not necessarily do it all yourself. But you will have to show your knowledge because you will work with machine learning engineers to make sure your model actually works.

Trouble:

• Understand the three main ways data processing: Party processing, in real and hybrid time.
• Understand Machine learning pipelinesIN CI/CDAND Monitoring the machine learning model.
• Exercise work flow in your projects, including Data consumptionmodel trainingIN VersionAND portion.
• Familiarize yourself with the tools of machine learning orchestration, such as Prefect AND Airflow (to orchestration), Kubeflow AND Zenml (for the abstraction of the pipeline) and MLFLOW AND Weights and prejudices (to track).

6. Without preparing for questions about a behavioral interview

Mistake: Rejecting questions such as “Tell me about the challenge you faced” as invalid and not preparing for them.

Why does it hurt: These questions are not part of the interview (only), because the interviewer is bored to death with her family life, so she would prefer to sit with you in a stuffy office, asking stupid questions. Behavioral questions check how you think and communicate.

Trouble:

7. Using fashionable words without context

Mistake: CV packing in fashionable technical and business words, but without specific examples.

Why does it hurt: Because “used the most modern synergies of large data sets to improve the scalable AI solution based on data for comprehensive cloud generative intelligence” really means nothing. You can accidentally impress someone. (But don’t count on it.) You will be more likely to explain what you mean and you risk that you have no idea what you are talking about.

Repair is:

• Avoid using fashionable words AND communicate clearly.
• You know what you are talking about. If you can’t avoid using fashionable words, then for every fashionable word, attach a sentence that shows how you used it and why.
• Don’t be unclear. Instead of saying “I have experience from DL”, say “I used Long short -term memory To forecast the demand for product and reduced supplies by 24%. ”

Application

Avoiding these seven mistakes is not tough. Making them can be high-priced, so don’t do them. The recruitment process in data science is sufficiently complicated and macabre. Try not to make your life even more complicated, thus succumbing to stupid mistakes as other data scientists.

Nate Rosidi He is a scientist of data and in the product strategy. He is also an analytical teacher and the founder of Stratascratch, platforms support scientists to prepare for interviews with real questions from the highest companies. Nate writes about the latest trends on the career market, gives intelligence advice, divides data projects and includes everything SQL.