Introducing f1facematch.com: End-to-End ML (Part 3/3)
After a five-year absence, the Chinese Grand Prix makes a welcome return this weekend to Formula 1. Just to paint a picture of how much F1 has changed since then, the podium of the 2019 Chinese GP was Lewis Hamilton in P1 followed by Valtteri Bottas and Sebastian Vettel. Fast forward to 2024, Lewis Hamilton has not won a race in two years, Bottas has zero points to his name after four races, and Sebastian Vettel is no longer racing. One of the most important qualities of a successful F1 driver is his ability to understand his car. To do that, he must understand the important work of hundreds of employees in the factory and the paddock that helps him drive that car as fast as possible during the weekend.
Working as a data professional is no different. You could be working as a data analyst/scientist/engineer or ML/software engineer, but to know what your teammates do in their job and how that helps you do your job is an important thing to learn. If you are a data scientist, you should understand the job of a data engineer in order to use the available data optimally and even suggest ways we can store data in the most useful way. Once the data comes to you, it’s important to keep ML/Software engineers in mind who will use your model to build ML pipelines and software. Below is an attempt by me to understand what it takes to build an ML tool from scratch.
Steps in the project
If you’re working in a somewhat large data team, you might have sub-teams dedicated to each section in the above workflow. But if you’re in a smaller team (sometimes even a data team of one), you might have to do all this on your own. Regardless of your situation, or the industry you’re in, it is hard to ignore the benefits of having end-to-end experience in the services you provide or the products you build. Whether you’re crunching numbers as a data scientist or tearing up the Formula 1 track, the value of hands-on experience spans industries and roles, enriching your career and broadening your horizons.
Server Architecture
However, I was not completely alone on this journey. Resources such as Youtube and ChatGPT helped me navigate a space that I’ve never been to before. I usually stop working once I’ve made a model in a Jupyter Notebook that is “good enough” for the job at hand. Anything beyond step 3 from the above process was unfamiliar territory for me. However, because of the vast amount of resources available to learn out there, through the work that people have done before me helped me create this end-to-end ML tool. Below are some resources that were extremely helpful for me to finish this project.
- Deploy machine learning model to production AWS (Amazon EC2 instance) by codebasics.
- HTML CSS and Javascript Website Design Tutorial — Beginner Project Fully Responsive by Brian Design.
- Debugging using ChatGPT
As the 2024 Chinese GP weekend kicks off, why not add a dash of fun with f1facematch.com? Discover which Formula 1 driver you resemble and maybe even surprise yourself! Test it out on your friends and family for a good laugh. Just remember, it’s all in good fun, and if your ‘twinsie’ isn’t your favorite driver, blame the model builder — not the messenger! Keep in mind, that even the most advanced machine-learning models have their quirks and imperfections. It’s all part of the journey towards smarter tech. If you’ve made it this far, thanks for joining me on this adventure! Stay tuned for more exciting stories and experiences ahead.
