Physics undergraduate @ NIT Agartala · AI/ML · Mathematics · Scientific Computing · Software

I'm a physics student exploring machine learning and intelligent systems from the mathematical and physical side. I enjoy learning by building things, implementing ideas from scratch, and digging into the why behind a model rather than treating it as a black box.

Currently, I'm especially interested in learning, memory, reasoning, continual learning, and neural architectures.

• Machine Learning — neural networks, optimization, representation learning
• AI Memory — persistent memory, retrieval, continual learning, learning dynamics
• Neural Architectures — understanding where Transformers work and where they struggle
• Neuromorphic Computing — brain-inspired computation and efficient AI hardware
• Physics × AI — using mathematical and physical ideas to think about computation and intelligence
• ML Research — building the foundations to eventually develop and test my own ideas

Mathematics → ML Fundamentals → Deep Learning → Research

Right now I'm strengthening my foundations in:

• Linear algebra
• Probability & statistics
• Calculus and optimization
• Python, NumPy & scientific computing
• Neural networks and deep learning
• Data analysis and visualization
• Modern AI tooling and LLM systems

The long-term goal is to move from using models → understanding models → designing models.

• Full-stack web development
• REST APIs & API integration
• Authentication, authorization & JWT/OAuth concepts
• WebSockets / real-time applications
• Responsive UI development
• Database design & ORM-based development
• Data analysis & scientific computing
• Neural networks & deep learning fundamentals
• Embeddings, retrieval & LLM application patterns
• OCR and document-processing workflows
• Git-based development & CI/CD

I prefer projects that make me learn something while building them.

Some of the things I've built or explored include:

• ML algorithms implemented from scratch
• Small neural networks and experiments
• AI systems with persistent memory
• Experiments with alternative neural architectures
• Scientific computing and physics simulations
• Full-stack web applications
• AI-powered tools and LLM applications
• Developer tools and automation projects
• Small products that turn technical ideas into useful software

Can memory be a fundamental part of an intelligent architecture rather than an external component?

How can a model learn continuously without catastrophic forgetting?

What kinds of architectures could complement or eventually replace today's dominant approaches?

Can ideas from physics and mathematics give us better ways to understand learning and intelligence?

Physics → Mathematics → Computing → Machine Learning → AI Research

My long-term goal is to work at the intersection of physics, mathematics, and AI, and eventually contribute to systems that can learn, remember, and reason in fundamentally better ways.

For now: learn deeply, build constantly, and follow the hard questions.