Hi 👋, I'm Mayank Singh

I’m a Computer Science student focused on AI and DevOPS.

I enjoy building things from scratch — neural networks, system-level tools, and full-stack applications — to deeply understand how they work under the hood.

Interests

• Machine Learning & Deep Learning
• Linux internals and tooling
• AI Systems & Optimization
• Cybersecurity and Penetration Testing

Research Work

Kuramoto–Sakaguchi Non-Local Equations (PINN)

• Studied Deep Neural Networks for solving non-local conservation laws derived from the Kuramoto model.
• Analysed the impact of activation functions, depth, width, and training setup on convergence and accuracy.
• Compared PINNs with classical numerical methods, highlighting accuracy trade-offs and limitations near discontinuities.

Biologically Plausible Learning for Medical AI

• Investigating alternatives to backpropagation using Direct Feedback Alignment (DFA).
• Introduced ViT-DFA, applying DFA-based learning to Vision Transformers.
• Evaluating convergence, robustness, and generalisation on medical image classification tasks.

Currently...

• Ricing my linux for programming efficiency
• Working on research-oriented ML projects
• Improving system design and engineering fundamentals

Tech Stack

Connect with me:

• GitHub: https://github.com/m4YnK-7
• LinkedIn: https://linkedin.com/in/mayank-singh-patwal