My colourful section

Here is a red highlight and a normal link to Next.js.

My Journey Learning Machine Learning

It's been almost a year since I started my journey into artificial intelligence and machine learning, and what a ride it's been! I wanted to share some reflections on the experience and what I've learned so far.

Starting from Zero

When I first started, terms like "CNN," "gradient descent," and "overfitting" were completely foreign to me. I remember struggling with basic concepts like understanding the difference between supervised and unsupervised learning.

My First Projects

Dog vs Cat Classification

My first serious ML project was building a CNN to classify images of dogs and cats. It sounds simple now, but at the time, getting TensorFlow to work and understanding how convolutional layers processed images felt like magic.

# My first CNN model (simplified)
model = Sequential([
    Conv2D(32, (3, 3), activation='relu', input_shape=(150, 150, 3)),
    MaxPooling2D(2, 2),
    Conv2D(64, (3, 3), activation='relu'),
    MaxPooling2D(2, 2),
    Flatten(),
    Dense(512, activation='relu'),
    Dense(1, activation='sigmoid')
])

Moving to Ensemble Methods

Later, I worked on a drinking/smoking prediction project using ensemble learning with LGBM, XGBoost, and Random Forest. This taught me that sometimes combining simpler models can outperform complex ones.

Kaggle Competitions

Participating in Kaggle competitions has been incredibly educational:

• Datathon 2023: Ranked 113/255 - Not amazing, but my first taste of competitive ML
• DTC Zoomcamp Q&A Challenge: Ranked 21/54 - Much better!

The key lesson? Feature engineering often matters more than fancy algorithms.

What I've Learned

Technical Skills

• Deep Learning: CNNs for image classification
• Traditional ML: Random Forest, XGBoost, LGBM
• Tools: Python, TensorFlow, scikit-learn, pandas
• Deployment: Basic model serving and APIs

Soft Skills

• Problem-solving: Breaking down complex problems into manageable pieces
• Research: Reading papers and implementing techniques
• Patience: ML projects take time, and debugging is part of the process

Current Focus

Right now, I'm particularly interested in:

• Hardware acceleration: Learning about custom ASIC and FPGA design for LLMs
• Production ML: How to deploy and maintain models in real-world applications
• Computer vision: Exploring more advanced CV techniques

Advice for Beginners

1. Start with projects, not just theory - Build something, even if it's simple
2. Join communities - Kaggle, Discord servers, university clubs
3. Don't get overwhelmed - The field is huge, focus on one area at a time
4. Practice regularly - Consistency beats intensity

What's Next?

I'm planning to dive deeper into:

• Advanced computer vision techniques
• MLOps and model deployment
• Contributing to open-source ML projects
• Maybe starting a YouTube channel to share what I learn?

The journey is far from over, but I'm excited about where it's leading!

---

What's your ML journey been like? Feel free to reach out and share your experiences!