Roshan Kumar Gupta

Success Story

Life Journey & Milestones

Biography & Background

My Journey My journey in technology began at SawRaj Institute, where I discovered that true education is not about memorizing answers but about learning how to think. SawRaj provided an environment that encouraged curiosity, creativity, and independent problem-solving. Rather than focusing solely on examinations, my mentors constantly challenged me to question assumptions, explore new ideas, and approach problems from first principles. That mindset became the foundation of everything I have built since. The confidence I gained at SawRaj gave me the courage to participate in hackathons, pursue ambitious research projects, and explore fields that initially seemed beyond my reach. Whether it was artificial intelligence, quantum computing, astronomy, systems engineering, or entrepreneurship, I never felt limited by what I knew—I was motivated by what I could learn. After joining Chandigarh University, I continued building on that foundation. I participated in over 15 national hackathons, winning competitions such as ZeroToOne and Zinnovatio, where I worked on innovative solutions in artificial intelligence, Web3, and software engineering. My passion for research led me to develop Project TARS, an AI-powered platform for exoplanet discovery using NASA's TESS mission data. The project later received support through a $5,500 Emergent Ventures Grant, enabling me to expand the research into larger-scale astronomical inference and scientific machine learning. Alongside TARS, I began working on EarthOS, a long-term research initiative exploring cognitive architectures for human-like machine intelligence. EarthOS investigates reasoning, memory, planning, uncertainty, and continuous learning with the goal of building AI systems that are more interpretable, trustworthy, and capable of genuine scientific reasoning. My research has since been presented at international conferences, including CCNCPS 2026 in Dubai, and I have authored publications in artificial intelligence and quantum computing. Beyond research, I have also gained industry experience as an L3 IT Intern at SMG Electric Scooters Ltd., where I developed enterprise software solutions and learned how research principles translate into real-world engineering. Looking back, every milestone—from hackathon victories and research grants to publications, internships, and ambitious research projects—can be traced back to the environment that SawRaj Institute created for me. SawRaj didn't simply prepare me for examinations. It taught me how to think independently, solve problems creatively, and believe that ambitious ideas are worth pursuing. Those lessons continue to guide every project I build, every paper I write, and every challenge I take on. I will always be grateful to SawRaj Institute for giving me not only knowledge, but the confidence and mindset to turn curiosity into innovation. Below are my Projects: EarthOS Status: Ongoing Research Initiative EarthOS is a next-generation cognitive architecture designed to explore how artificial intelligence can move beyond pattern prediction toward structured reasoning, persistent memory, planning, uncertainty estimation, and long-term knowledge acquisition. Unlike conventional AI systems that primarily generate responses from statistical patterns, EarthOS investigates how intelligent agents can build internal world models, learn continuously, and reason through complex multi-step problems. The project combines concepts from cognitive science, predictive processing, active inference, information theory, deterministic systems engineering, and modern machine learning. Its modular architecture includes dedicated components for memory management, planning, reasoning, knowledge representation, uncertainty modeling, tool use, and adaptive learning. EarthOS serves as a research platform supporting multiple publications and experiments while providing a foundation for future AI systems capable of scientific reasoning, autonomous problem solving, and trustworthy decision making. Technologies Python • PyTorch • CUDA • C++ • Information Theory • Active Inference • Cognitive Architectures • LLM Systems • Knowledge Graphs Project TARS AI-powered Exoplanet Discovery & Validation Platform Project TARS is a scientific machine learning platform developed to assist astronomers in identifying and validating exoplanets using observations from NASA's Transiting Exoplanet Survey Satellite (TESS). The system combines astrophysical transit modeling with machine learning to distinguish genuine planetary transits from stellar variability and false positives. Beyond classification, TARS incorporates physics-constrained inference, feature engineering, transit signal analysis, Bayesian validation techniques, and astronomical visualization tools to create a complete research workflow. The platform aims to accelerate exoplanet discovery while maintaining scientific interpretability. The project received support through an Emergent Ventures grant and continues to expand toward large-scale astronomical inference and planetary system analysis. Key Highlights Physics-constrained ML pipeline Transit light curve analysis Feature extraction using FFT Bayesian validation framework Scientific visualization NASA TESS dataset integration Technologies Python • TensorFlow • LightGBM • XGBoost • NumPy • Pandas • Lightkurve • Astronomy • Scientific Computing QUANTA Operational Observability in Noisy Quantum Systems QUANTA is a research project investigating the fundamental limits of extracting useful information from noisy intermediate-scale quantum (NISQ) computers. The research introduces the concept of the Entropic Noise Boundary, describing the operational horizon beyond which quantum information becomes increasingly difficult to observe despite remaining physically present. The project explores relationships between entropy growth, information degradation, quantum decoherence, and computational observability. It proposes mathematical models that characterize transition regions between useful computation and irrecoverable information loss, providing insights into practical limitations of current quantum hardware. This work contributes toward understanding scalable quantum computation from an information-theoretic perspective. Technologies Quantum Information Theory • Linear Algebra • Python • Scientific Computing • Mathematical Modeling GPU FieldChain High-Performance GPU Storage Integrity Framework GPU FieldChain is a high-performance storage integrity architecture designed to leverage modern GPUs for data verification and validation. Traditional storage integrity mechanisms often rely heavily on CPU resources, whereas FieldChain offloads computationally intensive integrity operations to massively parallel GPU pipelines. The system utilizes Vulkan compute shaders, custom memory layouts, optimized data pipelines, and deterministic execution strategies to achieve extremely high throughput while maintaining data consistency. Performance benchmarks demonstrated sequential write speeds of approximately 7 GB/s and bandwidth significantly exceeding specialized enterprise hardware solutions. The project explores the intersection of graphics hardware, storage systems, and high-performance computing. Technologies C++ • Vulkan • Compute Shaders • GPU Programming • PCIe 4.0 • Parallel Computing Roshan Arch Deterministic Runtime & Memory Architecture Roshan Arch is an experimental deterministic runtime architecture focused on eliminating unpredictable garbage collection pauses through explicit memory management and compact state representation. The architecture introduces efficient bit-packed state protocols, deterministic scheduling strategies, and optimized memory allocation techniques to improve runtime predictability for latency-sensitive systems. The design is particularly suited for robotics, real-time simulations, scientific computing, and embedded AI applications where deterministic execution is essential. The project explores alternatives to conventional managed runtimes while emphasizing performance, predictability, and low-level systems engineering. Technologies C++ • Runtime Systems • Memory Management • Operating Systems • Data Structures • Systems Programming HSA* Heuristic Safety-Aware Pathfinding Algorithm HSA* is an optimized pathfinding algorithm extending traditional graph search techniques by incorporating safety-based heuristics alongside distance optimization. Rather than minimizing travel distance alone, the algorithm dynamically evaluates environmental risk using inverse-square safety weighting to generate routes that better balance efficiency and safety. Performance optimizations reduced computation time to microsecond-level execution, making the algorithm suitable for robotics, autonomous navigation, drones, and intelligent transportation systems requiring rapid real-time path planning. Technologies Algorithms • Graph Theory • C++ • Computational Geometry • Autonomous Navigation Cryptic DAO Treasury Dashboard Web3 Financial Management Platform Cryptic is a modern treasury management dashboard developed during the ZeroToOne Hackathon for decentralized autonomous organizations (DAOs). The platform provides a unified interface for monitoring treasury assets, transaction history, governance activity, and capital allocation across blockchain ecosystems. As Frontend Lead and UI/UX Designer, I designed a scalable component architecture, interactive dashboards, responsive layouts, and data visualization systems focused on clarity, usability, and enterprise-grade user experience. Technologies React • TypeScript • Next.js • Tailwind CSS • Figma • Framer Motion • Web3 Internal CRM Portal – SMG Electric Scooters Industry Internship Project During my internship at SMG Electric Scooters Ltd., I designed and developed an internal Customer Relationship Management (CRM) portal to streamline logistics, vendor management, and operational workflows. The platform centralized data management, automated routine processes, and improved communication between departments, reducing manual effort while enhancing operational visibility. The project provided practical experience in enterprise software development, database design, and workflow automation within a manufacturing environment. Technologies React • Node.js • SQL • REST APIs • Enterprise Software Development