📻 😹 👨🏼‍🎨 How a beginner Data Scientist like me can gain experience 🤲🏼 👶🏻 👩🏼‍🎨

I have completed several courses, but what should I do next?

The other day I received a letter from Manuela with the following question: “How can a novice data scientist like me gain experience?”

Great question.

The question followed: “I completed several courses and projects, some employers said that they like my education, but I lack the minimum experience, and so I would write to them after receiving it!”

In my answer, I have described several points that I will copy into this article.

Disclaimer: I'm not looking for a job. You might be asking, "Why should I even listen to advice from someone who's not looking for a job?"

And you will be right.

Always be skeptical of advice to those who have not learned from their own experience. And here's my explanation: none of the work I found suited me, so I created my own (you can write a separate article on this topic, but here's my advice - don't look for a job, it's better to start your own business ).

However, if I were still looking for work (and, in turn, the experience required for it), I would do the following things.

Gap between courses and real experience

Online courses are consumer goods (very common). Yes, I say this as someone who conducts an online machine learning course himself .

The online course forms fundamental knowledge, independent work on projects - specific knowledge (those that cannot be taught).

Therefore, my first question to Manuele is: “When you say that you have completed several courses and projects, how many of the projects were your own and how many were part of the course?”.

I define experience as follows: things that you have tried on your own (or in collaboration with others), without knowing the result.

Let me give you an example.

How many times have your parents told you that the stove is hot?

I believe that you did not believe until you convinced yourself.

Listening to parents is the same as listening to the teachers of the course. Of course, you can accept what they say, but without your own experience, you will not internalize this knowledge.

So getting experience (before starting work) is equivalent to touching the stove.

If you, like me, have certificates and course projects, even chew your ass, and you are looking for this elusive thing called “experience”, now is the time to ask the question:

Have I touched the stove yet?

Start working before you get the job

Find the perfect job that you would like to do, and then start doing it.

If this sounds crazy to you, then just treat it as part of your research. If you were able to understand data science, then you can understand who exactly is required in the labor market.

Let me give you a concrete example.

Let's say you enjoy freelancing, but experience is required everywhere. And there is one job in which you need to build a machine vision model in order to understand whether the doctor's instruments were on the operating table (I did not invent this, this is from a real project that I was asked to work on).

You read the duties and think "wow, but this is interesting, I would like to work on it." You read on and see that 3+ years of work experience is required and you think “so, I don’t have such experience, so it’s not worth submitting.” And then you torment yourself all day with the thought that you are not good enough for any job.

I understand you perfectly. This is the wrong thought process. This is a decision based on fear, not growth. You decide not to apply for this job because you fear that your abilities are not up to par.

The truth is, almost everything has to be figured out on the fly (after all, this is a science in "data science" or an engineer in a "machine learning engineer").

So what can you do?

You can take a description of the project and make it yourself. Spend a week developing and planning the process, four weeks on the process itself, and then share what you've learned for a week.

In the worst case, you will spend six weeks on something that does not work (for a particular situation), in the best case, you will get what you can show, what you did yourself (without a course project guide).

You will be able to say: “I saw this project, took up it and decided to confirm the concept's efficiency”

In the situation with the doctor’s tools, you can search the Internet for photos of the tools, create your own data set (it doesn’t matter if it is not real, it’s important to understand how to create it), build a model that determines whether the tools are in place, and then turn the concept check into a good one application using Streamlit .

An example of what a stand-alone project might look like. Start by collecting data, model or analyze it, create a user interface with Streamlit (or any equivalent) and put it into action. For a complete example, check out how I copied Airbnb amenities using the same steps.

If your future employer cannot call such an initiative “experience”, this work is not suitable for you.

Note: When Andrei (my business partner) and I hired a teaching assistant for our machine learning course (hi, Shubhamai), we chose a student who already independently answered the questions of others. He started working even before he got the job.

Follow up

That which is always neglected. Never take failure as a constant that never changes.

As already mentioned, some employers wrote: "you do not have the desired experience, write to us how you get it."

Let's change one word: "you do not have the desired experience yet, write to us how you get it."

If someone tells you that you have no experience yet, and you are conscientious about their words and practice your skills, namely, starting work before getting the job itself, you should re-write to these people.

Show them what you were working on (it should be noted that this requires really improving your skills).

Write to them: “I wanted to let you know that since our last conversation I have improved my skills and this is what I have done during this time. If you have any open positions, you may want to consider me again. "

Many forget about this step. And this applies not only to interviews, but to anything else. Many relationships can be improved if someone takes the extra step in that direction.

Remember, many results in life are non-linear. If I forget to reply to an email within a couple of months (email is my best friend and worst enemy), then it is likely that I will never reply. But if this person re-writes to me and reminds me of his own laxity (thank you), most likely I will answer within 24 hours.

Does it always work?

Not. Of course not. But this way you show more initiative than simply responding to several vacancies.

Meet people

Many open positions are not even published.

I got my first (and only) job at a tech company because someone saw my posts on LinkedIn, who knew someone who knew someone who might be interested in me from a professional point of view. I ended up meeting this person (hello Cam) who turned out to be a living legend. We talked about my work, he asked if I would like to come for a day next week - I came, then came again, and a couple of days later I was offered a job.

The important thing in this story is that I had fundamental knowledge, of course they did not pull to the highest level, but I knew what pandas DataFrame was or how to troubleshoot and search for questions on Stack Overflow (yes, I spent a lot of time studying and just googling all sorts of things; remember, an engineer is someone who "figures things out").

Many of our employees (especially the younger ones) got a job because someone knew someone.

Here's what I think: job portals are dead.

Perhaps they still help some people, but certainly not in my world. I convince myself that it’s easier to respond to work with one button than to try my best, find a person to interview, talk to him, show the fruits of his work, listen to his advice, feedback, improve his skills, contact him again, and, maybe get a job offer. While if I respond to a vacancy through the site, for me it is an instant refusal (again, maybe this is not true, but I see it that way).

Of course, in an ideal world (although perhaps not ideal) companies would only hire people for their talent and abilities, but this does not happen. People like to work with people they like.

Okay, so how do you get to know someone?

If you chose the digital route, keep it simple. Just like you write to me (you saw that Maneula sent me a very concise and specific letter). Briefly and to the point. Be straightforward.

I got an appointment on Airbnb because I sent a recruiter (whom I was specifically looking for) a LinkedIn message. She answered me during the day: “Hi Daniel, is it convenient for you to call tomorrow?”. I did not get a job, but I was definitely closer than if I simply responded through the site (which I didn’t even do).

Again, success is not guaranteed.

You can write to a hundred people and not get a single answer. In the same way as if you had responded to a hundred vacancies and would not have received a single one. And if in the end it turns out that way, think of it as a feedback - you need to improve your approach to things.

Digital is efficient, teleworking is becoming more and more popular, but people still enjoy interacting (surprisingly for tech nerds like us) with each other, being in the same room. And this means attending events (when social conditions allow), talking with people, taking extra effort to ask Dave from Company X about his project, on which he worked over the past few months.

The company I worked for holds a meeting on artificial intelligence every month. I went to one such meeting without knowing anything about the company, and after seven months I worked in it.

Share your work

It seems to me that everything is clear with this, but I'd rather write about it again.

If your attempts to seize opportunities fail, have them come back.

It starts with creating things and sharing them publicly (no matter if they're finished).

Over the past three years, I have been getting great opportunities because someone has come across my work (I have been writing in the online space since 2016, from 2017 about machine learning).

Yes, create your own blog and use it as the basis of your portfolio, and your other social networks (you will understand which ones) as support for the main platform.

Think about it. How did you learn something? Most likely you were looking for material, stumbled upon an excellent blog post and used it in your work.

And you know what?

You yourself can create such valuable blog posts.

This connects us with what I have said. Start working before you get a job and share what you learn (your experience). And as soon as you work on something related to the previous refusal, write to them again and say: “And here I am.”

How a beginner Data Scientist like me can gain experience