A is the new resume — it’s what substitutes for real work experience.
But right now, your projects are either useless filler or you’re simply not taking them seriously, and that’s why you’re not landing interviews.
So in this , I’ll break down the essential project types that top-tier companies actually look for, so you can stop submitting dead-end applications and start scheduling interviews.
Let’s make your portfolio the interview magnet it needs to be.
3-5 Simple Projects
The absolute baseline for your portfolio is 3–5 “simple” or “easy” projects.
This will not necessarily move the needle in getting hired, but it will give your portfolio initial weight.
Think of these simple projects as the “warm-up reps” at the gym. They aren’t the heavy lifting that builds serious muscle, but they establish the fundamental mechanics, consistency, and discipline needed before you tackle the main challenge.
The primary goal of these projects is to get you creating and building without a guided tutorial, and to really get you thinking creatively about how to solve problems.
It’s also about “optics” and ensuring that your resume, GitHub, and LinkedIn profiles appear active and well-populated.
However, do take about a month to build these smaller projects, ensuring they are of sufficient quality and not hastily generated with ChatGPT.
Aim to build a wide range of projects, each using different tools, datasets, and machine learning algorithms.
If you want some inspiration, check out this repo I made nearly 5 years ago, which contains examples of these simple projects when I was trying to get my first job.
GitHub – egorhowell/Data-Science-Projects: A selection of small Data Science Projects.
A selection of small Data Science Projects. Contribute to egorhowell/Data-Science-Projects development by creating an…github.com
One thing I will say is that these projects are probably below par by today’s standards, as the field is becoming increasingly competitive.
So, below is a list of key objectives that your simple projects should meet to make them worthwhile:
- Variety of Algorithms — try to include Gradient Boosted Trees, Neural Networks, and clustering algorithms like K-Means and DBSCAN in your projects.
- Novel Data — It’s much better to obtain a messier and more realistic dataset that reflects the data you will encounter in the real world. This will impress employers and interviewers even more, directly demonstrating your data science and machine learning skills.
- Personal — To decide what your projects should be on, it’s best to start by answering specific questions you think will be interesting to discover from the data. A personal touch is always better.
End-To-End Project
If you want to work in machine learning, you need to be able to deploy your algorithm.
“A model in a Jupyter notebook has zero business value”
You have probably heard this sentence from me and others multiple times.
Having the most sophisticated, fanciest state-of-the-art transformer model means absolutely nothing unless it is making real-life decisions.
Companies and hiring managers know this, and frankly, all they care about is whether your model is saving or making them money and whether their underlying profit is increasing.
It is really that reductive.
So, you want to showcase to potential employers that you know how to build and ship an algorithm end-to-end in your portfolio.
Your project should ideally include the following:
- Data collection and storage.
- Data preprocessing.
- Model training and evaluation.
- Model deployment (via API, web app, VPS, etc).
- Analysis and presentation of your results.
This project is often the hardest for beginners to create because it does require some up-skilling and learning a bit of software engineering.
Some of the things you will need to learn are:
What I don’t want you to do is get intimidated and overwhelmed by the list.
Start small and learn the essentials as you go; you certainly will not need to use everything I just mentioned.
And as always, make it as personal as possible; this will keep you motivated, and it’s a much better talking point in interviews.
If you want a real-life example, then check out one of my previous YouTube videos where I walk through a complete end-to-end project I created that forecasts stock prices and then optimises my portfolio.
Research-Focussed Project
I often recommend that people add some research element to their portfolio.
One method is to re-implement a research paper they are interested in.
You will learn so much from this process:
- Understand complex maths associated with cutting-edge models.
- Implement sophisticated models from scratch or using simple libraries.
- To think creatively and apply your own knowledge to new ideas.
- Improve your understanding of current trends in the field and what top researchers are working on.
And the best part is that the majority, literally 99%, of candidates are not doing this, so you will instantly stand out.
Some useful websites to find papers:
Re-implementing a paper is very hard. I have tried several times in the past, and I still couldn’t quite get it 100% correct, but I learned so much from that process.
Another way to add research into your portfolio is through reading and distilling papers either through writing about it online or even through a journal club.
The latter is what I set up at my previous company, and it was beneficial. I presented a variety of papers such as:
It taught me how to translate some of the most technical topics in the world at the moment into a digestible 1-hour presentation.
This is a skill that companies really desire, as many practitioners in the field don’t have it.
If you currently don’t work in a company where you can set up something like this, there are many Discord and community groups out there.
One group I recommend is Yannic Kilcher’s Discord. He is a machine learning researcher and engineer who creates YouTube videos breaking down research papers.
Write Technical Articles
Most people assume their articles need to be “groundbreaking.”
What if I told you that’s just an excuse, and your blog doesn’t need to be unique to land you a job?
If you look at mine, most of the posts are about fundamental statistical, data science and machine learning concepts.
To date, I have written well over 150 technical and over 60 career-based advice articles.
These started off purely for myself as a means to learn more about the field; I didn’t care if people liked them or not, as they were solely for me.
This is the attitude you should have as well.
Start by documenting what you are currently learning or want to learn. No need to overcomplicate it.
Having a blog brings so many positives to your career and abilities:
- Solidifies understanding of concepts.
- Helps you think and have better communication skills.
- Demonstrates a self-starter attitude and an interest in the field.
- Will literally land you jobs and interviews. This happened to me!
Your blog is a passive income generator for your career. The earlier you invest in it, the better the payoff.
I recommend you start blogging here on Towards Data Science, as it’s very easy to use, has a large data science community, and already has an in-built audience.
There are other, more developer-focused platforms, such as Hashnode, or you can even blog on your own website, utilising platforms such as WordPress or Ghost.
You can even have your own blog that you create from scratch using HTML, CSS and JavaScript!
If you want to learn more, I have a whole post about how to start and write a technical blog that you can check out below:
Now that you know the exact projects that turn your portfolio into an interview magnet, there’s just one final piece of the puzzle: how you present it.
Most people just throw a GitHub link on their resume and hope for the best, but if you do that, you’re missing out on a huge opportunity to highlight the business value of your work.
To learn exactly how to showcase your portfolio, see one of my previous posts below.
I will see you there!
Another Thing!
Join my free newsletter where I share weekly tips, insights, and advice from my experience as a practising data scientist and machine learning engineer. Plus, as a subscriber, you’ll get my FREE Resume Template!
Dishing The Data
main pagenewsletter.egorhowell.com