🧠Catching Up with AI: What I Learned from My First ML & DL Cohort as a Student
Unless you’ve been living completely offline, you’ve probably seen how AI and Machine Learning are reshaping the future. From ChatGPT to self-driving cars, the hype is real — and as a student, I felt the urge to catch up.
That’s when I joined a student-friendly Machine Learning & Deep Learning cohort, and it became one of the best learning experiences of my tech journey so far.
🚀 Why I Joined the ML/DL Cohort
I’ve always been more focused on frontend development — HTML, CSS, JavaScript, React — but I didn’t want to ignore the growing role of AI. So when I saw a chance to explore Machine Learning, I grabbed it.
The cohort was designed for beginners: structured lessons, hands-on exercises, and clear explanations. No math-heavy fear, just curiosity and code.
đź“š What I Learned
Here’s a breakdown of the concepts and topics we explored:
🔹 Machine Learning Basics
- ML is about training a model using data to make predictions or generate output.
- There are 3 main types of ML:
- Supervised Learning (with labeled data)
- Unsupervised Learning (without labels)
- Reinforcement Learning (with rewards/penalties)
🔹 Core Concepts in Supervised Learning
- Features & Labels: Features are input data; labels are the output we want to predict.
- Model Training: The model adjusts its parameters based on the difference between its predictions and the actual label (aka loss).
- Evaluation: After training, the model is tested using data it hasn’t seen before.
🔹 Regression & Classification
- Regression predicts a numeric value (e.g., predicting rainfall).
- Classification predicts categories (e.g., spam or not spam).
🔹 Gradient Descent & Loss
- We use loss functions (like MAE, MSE) to measure how wrong predictions are.
- Gradient descent is used to minimize this loss by updating weights and biases.
🧠What About Deep Learning?
We also explored Deep Learning, a powerful subset of ML inspired by the human brain.
🔸 Neural Networks
- Neural networks are built with layers of neurons.
- Each neuron processes inputs with weights and biases, then passes it to the next layer.
- Activation functions decide whether a neuron should "fire" or not.
🔸 Types of Neural Networks
- Convolutional Neural Networks (CNNs) – great for image classification.
- LSTMs – ideal for sequence data like speech or text.
đź› Tools We Used
- Python – the language we used for almost everything
- Pandas & NumPy – for handling and cleaning data
- Matplotlib & Seaborn – for data visualization
- Scikit-learn – for training models like Linear Regression, Decision Trees
- Google Colab – to run code in the cloud without setup
đź’ˇ My Takeaways as a Beginner
- Start small – Learn the concepts, not just the syntax.
- Practice more than you watch – Run code, tweak parameters, see what changes.
- Visualization helps – Graphs and charts made everything more understandable.
- Don't fear the math – You can still learn the logic first and go deeper later.
- ML can fit anywhere – Even my digital journal app might use mood-based ML soon!
🌱 What’s Next?
While my focus remains on frontend and UI/UX, I now see how ML can complement my work. Whether it’s a recommendation engine, mood tracker, or smart health alert — I’ve opened the door to AI-powered features in my future projects.
🙌 Final Thoughts
To anyone who’s scared to get started with ML or feels like it’s “too advanced” — don’t worry. I started with no confidence in this domain, and now I’m genuinely excited to keep learning.
If you’ve just started or are curious about ML/DL, let’s connect! Drop a comment or DM me — always happy to chat with fellow learners!
Thanks for reading! 🚀
Top comments (0)