Loading…
Teaching
I believe teaching is one of the best ways to deepen understanding. Here is my course experience and pedagogical reflection at Carnegie Mellon University.
Supporting CMU's 11-785 Introduction to Deep Learning course through student mentoring, technical discussions, and assignment development. Contributing to course infrastructure by validating deep learning implementations, running model training and evaluation checks, and improving reference workflows across multiple architectures.
I believe in learning by doing. My approach focuses on helping students build intuition through hands-on projects and guided exploration of real-world problems in deep learning and machine learning.
Whether it’s discussing attention mechanisms or debugging transformer implementations, I aim to make complex topics accessible while maintaining intellectual rigor.
Students benefit most when tensor operations and architectural choices are grounded in mental models before jumping straight into syntax or heavy mathematical formulas.
Rather than hiding errors, walking through real stack traces, gradient explosions, and loss curves teaches the diagnostic instincts that textbooks rarely cover.
Introducing concepts incrementally — from raw matrix math to full multi-head attention loops — lets students build confidence at each step of abstraction.
Office hours and technical discussions center on student-driven questions and collaborative problem-solving: diagnosing experiment failures together.