Weekly Highlights

1. Machine Learning Course
2. Marketing Attribution Model
3. People who Inspired me

Machine Learning Course

Hooray! I finally finished the Google Machine Learning course. It took me a month to get through the seven extensive modules, and I learned a ton, including:

• Google's generative AI capabilities
• Building ML models with AutoML and BigQueryML
• Vertex AI capabilities
• Feature engineering
• TensorFlow
• End-to-end machine learning pipelines and MLOps

The course was incredibly comprehensive, offering a detailed overview of the entire Google Cloud Platform for machine learning. However, I was a bit frustrated with how the UI functions. With so many options, it's difficult to remember everything, and the user interface isn't as developed as I'd hoped. For example, I tried to build a workflow with an Airflow instance and was fully dependent on the UI, which made it very difficult to navigate through errors. In the end, I built the Airflow instance locally, and it was much easier to debug and implement. While GCP might be more useful for larger projects, it feels like overkill for me right now.

Now that the course is complete, I'm shifting my focus to the certification exam. I bought a course with six sets of mock questions and will be practicing them all week.

Marketing Attribution Model

Last week, I shared that I successfully created a Docker Compose file and used an Airflow instance to automate my workflow. It was a challenging experience! At first, I was stuck on dependency issues, unable to install packages due to version conflicts. It took me two days to figure out the whole process, but I finally solved it by creating a virtual environment and using a Dockerfile to build the Airflow instance. The feeling of seeing my Python script finally run successfully was amazing.

Here are some key learnings from this project:

• Always create a requirements.txt file to keep your project environment separate and safe from system-specific environments.
• Don't fully rely on LLMs; look at the issue carefully and ask the LLM to walk you through its reasoning instead of just giving you a flood of information.
• Take a break when you've been stuck for hours. Stepping away from the problem and getting outside can help your brain relax and refresh.
• Document your troubleshooting process so you can refer back to it the next time you start a similar project.

You can find the complete end-to-end guide for this project here.

People Who Inspired Me

Even though I felt a bit distracted this week, I still found inspiration from the people I talked to.

I met some new people who are doing incredible work in the tech world. The more I talk to them, the more I realize how much I still have to learn and how much I want to grow my skills.

I also have to give a shout-out to the Supabase + Next.js ULTIMATE Auth Combo video. The creator literally taught me how to build authentication in a single video. What was even more amazing was that they provided all the scripts and full access to the GitHub repository so viewers could easily replicate the project. Amazing!

That’s all for this week. See you in the next one!