Research

Research & Publications

Internships, thesis work, and peer-reviewed contributions.

Publications

📄 IEEE Publication

Zero-Day Attack Detection Using Machine Learning

Proposed a novel ML-based framework for detecting zero-day cyberattacks in network traffic. The system uses ensemble classifiers combined with unsupervised anomaly scoring to identify previously unseen attack patterns without requiring signature databases. Evaluated on benchmark intrusion datasets with significant improvements in detection rate over state-of-the-art methods.

IEEE Proceedings Ensemble Learning Anomaly Detection Network Security Python

View on GitHub

📄

M.Tech Thesis

🔬

Ongoing · IIT Indore

LLM-Driven Intelligent Network Slicing Using O-RAN RIC

This thesis develops an autonomous network slicing framework for 5G/6G systems using Large Language Models deployed on the O-RAN RAN Intelligent Controller (RIC). xApps and rApps interface with the Near-RT and Non-RT RIC layers to enable real-time, context-aware radio resource management. The LLM interprets telemetry, SLA requirements, and traffic patterns to make dynamic slicing decisions — replacing hand-crafted heuristics with intelligent, adaptive policies.

Framework

O-RAN Alliance RIC · OpenAirInterface · srsRAN

AI Component

LLM-based policy engine · RAG pipeline · xApp SDK

Target

5G NR · PDCP/RLC slicing · QoS enforcement

O-RAN RIC xApp / rApp 5G NR LLM Python Network Slicing OpenAirInterface

Research Internship

🧬

Research Intern · IIT Kanpur

Neural Modeling of Genetic Circuits

Mentor: Prof. Abhilash Patel, IIT Kanpur

Developed neural network models to simulate and predict the dynamics of synthetic genetic circuits. Projects included modeling first-order and second-order biological system behaviors (gene regulation, toggle switches, oscillators) using neural approximations. Implemented cost functions for stability and response accuracy, and fine-tuned architectures for fast convergence. Explored learning-based control systems and their applications in accelerating biological circuit design — a key challenge in synthetic biology.

Modeled first and second-order biological system behaviors (gene regulators, toggle switches) using neural ODEs

Implemented RL-based controllers for stability optimization of genetic circuit dynamics

Designed custom cost functions for time-response accuracy and convergence in neural approximations

Contributed to understanding how ML can accelerate synthetic biology circuit design workflows

Neural ODEs Reinforcement Learning PyTorch Synthetic Biology IIT Kanpur