Undergraduate Student & Budding Researcher.

My interests span three areas I find persistently compelling, and increasingly, the territory where they meet:

Systems Design  |  How architectural decisions propagate through large-scale infrastructure. I am particularly interested in fault models, emergent failure modes, and the principled tradeoffs that separate systems that degrade gracefully from those that do not.

MLOps  |  The theory-to-production gap in machine learning remains poorly understood and routinely underestimated. I am drawn to the operational and engineering problems that arise when statistical models are embedded in live systems — reliability, reproducibility, drift, and the organisational conditions that determine whether any of this is possible at all.

Computer Vision  |  Representation learning for visual data, with a growing interest in how robustness and generalisability are achieved (or claimed to be achieved) in practice, and what the gap between benchmark performance and real-world deployment reveals about the state of the field.

Vicinal is runtime security infrastructure for AI systems. It sits adjacent to your AI calls, observes them in context, and intervenes only when something is about to go wrong. Vicinal provides a guardian layer for developers who already use a lot of AI and know how fragile it is. It evaluates user input before calling an LLM and protects against prompt injection, data exfiltration, and restricted content requests.

Python   Deep Learning   Guardrails on AI

AEGIS-NET is an AI-powered system that analyzes aerial imagery from UAVs to identify safe landing zones in real-time. This project uses the WildUAV dataset, a large-scale benchmark for monocular depth estimation in unstructured outdoor environments captured from UAV perspectives.

Drone Tech   Defense Systems   Applied ML

Wydey is a progressive web application designed to map mathematical functions to geometric visualizations and audio samples. By translating abstract equations into visual formats and pairing them with instrument sounds (like Piano, Violin, Synth, and Flute), Wydey makes exploring mathematics a multisensory experience.

Applied Mathematics   Vite   Java Script & CSS

• Deepening my understanding of MLOps tooling — experiment tracking, model registries, deployment patterns
• Reading across the systems and distributed computing literature
• Open to research collaboration, particularly on projects involving signal processing, computer vision, or ML system design

If you are working on something at the intersection of these areas and are looking for a collaborator, I am glad to hear about it.