⚡ Dissertation: TokenGuard, an adaptive orchestration layer for cost-efficient LLM agent systems combining semantic caching, complexity-aware model routing, and adaptive plan reuse, with a reproducible cost-latency-quality benchmark.

⚡ Relevant coursework: Advanced Deep Learning, Statistical Learning Theory, Convex Optimization, Information Theory.

⚡ Focus on efficient and agentic LLM systems, and their evaluation under realistic deployment constraints.

⚡ Final Year Project: Intelligent Reconfigurable Surfaces for UAV Communications, funded by NESCOM.

⚡ Sole undergraduate recipient of NESCOM Research Funding in the department.

⚡ Specialised coursework in AI, machine learning, signal processing, and wireless communications.