B.Tech Computer Science & Engineering · Barak Valley Engineering College (ASTU) · Silchar, India
I study the algorithms that discover algorithms. My research focuses on symbolic regression and evolutionary computation — specifically how we can constrain learning systems to produce mathematically interpretable, physically valid models.
Rather than treating AI as a black box, I'm interested in how representation, constraints, and search strategies fundamentally change the behavior of learning systems. Currently, I'm analyzing search-space bottlenecks in grammar-constrained evolutionary systems and building reproducible computational pipelines for automated scientific discovery.
|
A reproducible framework for studying grammar-constrained symbolic regression through validity-aware structural tracking and multi-trial empirical evaluation. |
Quantifies synergy and redundancy between search-space constraints in symbolic regression via grammar-based AST generation and Monte Carlo density estimation. |
|
A production-oriented fleet booking platform with PostgreSQL, Supabase, and Upstash Redis — secure auth, Row-Level Security, and scalable booking workflows. |
A multi-agent pipeline using Claude, Gemini, and Tavily to transform unstructured conversational data into structured, publishable narratives. |
EML Framework: Symbolic Regression Representation Study (2025 · Zenodo Preprint) doi.org/10.5281/zenodo.19991771
Randomness in Quantum Cryptography (2024 · Zenodo Preprint) doi.org/10.5281/zenodo.15867370
- Constraint-interaction effects in grammar-guided symbolic regression
- Expanding the EML Framework for reproducible symbolic regression experiments
- Improving the architecture and scalability of Pather Saathi
