ANUP DAS

A comprehensive cloud infrastructure automation toolkit built using Terraform and AWS. CloudForge streamlines the provisioning, deployment, and management of production-ready cloud environments with built-in best practices for security, monitoring, and reliability.

Simulated real-world production failures on AWS infrastructure to develop incident response capabilities. Built a full-stack web application environment, then systematically introduced failures including instance crashes, network partitions, and database corruption to test recovery mechanisms.

Designed and deployed a production-grade three-tier architecture on AWS using Terraform. The deployment includes a web tier, application tier, and database tier with proper network isolation, load balancing, and auto-scaling configurations.

Implemented a secure VPC architecture with bastion host access pattern on AWS. Designed layered network security with public and private subnets, NAT gateways, and strict security group rules to ensure minimal attack surface while enabling SSH access to private resources.

Comparative analysis of quantum-inspired and classical algorithms for solving the Travelling Salesman Problem (TSP). Benchmarked QAOA and quantum annealing approaches against classical heuristics to evaluate computational advantages across different problem sizes.

Developed a novel quantum-inspired heuristic algorithm for solving the Travelling Salesman Problem. The approach leverages quantum computing principles — superposition and interference — adapted for classical hardware, achieving competitive solutions for combinatorial optimization problems.

Authors: A. Das, R. Tekcham, K. E. Patton and N. Marchang
DOI: 10.1109/GCON65540.2025.11173289

Abstract: In this paper, a quantum-inspired optimization algorithm, Quantum Inspired Hazelnut Tree Search (QIHTS), is proposed to represent a modification of the classical Hazelnut Tree Search (HTS) approach. The promising nature of the hybrid optimization technique has been demonstrated through experiments in which the classical heuristic is mixed with quantum-inspired concepts to achieve an advantageous and scalable optimization technique applicable in real-life circumstances. One of the main objectives of QIHTS is to find a sufficiently good balance between exploration and exploitation by transforming the classical HTS concept into a quantum-inspired approach that employs superposition-based exploration, controlled chaotic dynamics, and probabilistic blending. We compare the HTS with the proposed QIHTS method using a range of benchmark functions, including unimodal, multimodal, composite, and shifted functions. All population settings, the number of generations, as well as the chaos function for the structured noise, can be guaranteed to have consistent characteristics due to the control of the experimental design. Through a range of benchmark tests, it has been observed that QIHTS outdoes its classical counterpart in terms of the rate at which it arrives at an optimal location and the quality of the solution. By integrating quantum-inspired methods, the proposed algorithm can cope with complex search spaces and at the same time bypass local optima much more effectively.