Research & Publications

My Journey into AI Research

Embarking on a research career in AI has been the most transformative experience of my academic life. In this article, I share the milestones, challenges, and lessons learned along the way.

How It Started

During my sophomore year, I took a course on "Introduction to Artificial Intelligence" that changed my perspective. The concept of neural networks fascinated me – how simple mathematical operations could mimic the human brain. I started with small projects, like building a digit classifier using MNIST, and gradually moved to more complex problems.

Key Research Areas I've Explored

• Natural Language Processing: Sentiment analysis, transformer models (BERT, GPT).
• Computer Vision: Image classification, object detection with YOLO.
• Reinforcement Learning: Q-learning, policy gradients for game playing.

First Research Paper: Optimizing Database Queries with ML

My first serious attempt at research was a project titled "Optimizing Database Queries for E-Commerce Platforms." I proposed using reinforcement learning to dynamically select indexes based on query patterns. While the results were modest, the experience taught me:

• How to frame a research question
• The importance of baseline comparisons
• How to write a literature review

Challenges Faced

Research is not linear. I faced countless dead ends, from model convergence issues to dataset biases. The biggest challenge was reproducibility – ensuring that my experiments could be replicated by others. I learned to document everything meticulously.

Resources That Shaped My Thinking

• Books: "The Master Algorithm" by Pedro Domingos, "Artificial Intelligence: A Modern Approach" by Russell & Norvig.
• Courses: CS229 (Stanford), Deep Learning Specialization (deeplearning.ai).
• Conferences: I regularly follow NeurIPS, ICML, and ACL proceedings on arXiv.

Advice for Aspiring Researchers

If you're considering a research path:

1. Start small: Reproduce existing papers before attempting novel ideas.
2. Find a mentor: Guidance from experienced researchers is invaluable.
3. Write constantly: Maintain a research log – it helps clarify your thoughts.
4. Embrace failure: Every failed experiment is a step toward understanding.

My journey is far from over. I'm currently exploring few-shot learning for low-resource languages and hope to contribute to making AI more accessible. If you're also passionate about AI research, feel free to reach out – I'd love to connect!