JunKyu Lee · Research Fellow · University of Essex
I work across cybersecurity, efficient AI computing, and hardware–software co-design. My research aims to make advanced computing systems more trustworthy, resource-efficient, and deployable beyond large-scale datacentres.
Research vision
“Future computing systems should not force a choice between security, efficiency, and practical deployment. The central challenge is to design them together.”
Research themes
SECT connects algorithmic ideas, systems architecture, and hardware-aware optimisation. The goal is not simply to produce accurate models, but to create complete systems that remain secure, measurable, and efficient.
Methods for trustworthy AI, software supply-chain security, formal security reasoning, attack-surface analysis, and secure-by-design system architectures.
Communication-aware inference for distributed and edge platforms, including activation compression, tensor parallelism, and resource-aware deployment.
Cross-layer methods linking architecture, algorithms, arithmetic, memory, communication, and security constraints.
Models and design techniques for understanding where energy is spent and how system-level decisions reshape performance, power, and reliability.
Current directions
These summaries describe active or developing research directions rather than a complete publication list.
Reducing activation-communication overhead in distributed inference through layer-wise vector quantisation, calibration, and low-bit representations for bandwidth-constrained platforms.
Studying structured perturbations across inputs, parameters, activations, prompts, and tool-use states to quantify stability and identify security-sensitive operating regions.
Runtime and architectural techniques that coordinate memory pressure, data movement, replication, and scheduling in resource-constrained accelerator clusters.
Analytical techniques connecting arithmetic, data movement, voltage, throughput, and architectural constraints to support transparent efficiency comparisons.
How SECT works
Teaching
My teaching connects foundational programming and mathematics to the way real computing systems are designed, evaluated, and improved.
Python programming, problem decomposition, functions, data structures, files, exceptions, recursion, and object-oriented programming.
Quantitative foundations for scientific reasoning, data analysis, and later study in computing and engineering.
Assessment, lab support, technical feedback, and the development of practical problem-solving skills.
Career
My work combines numerical analysis, SW/HW engineering, AI applications, and cybersecurity.
Research and teaching in AI systems, efficient computing, and cybersecurity.
Research in computing systems and hardware–software co-design.
Research experience in FPGA-based machine learning systems.
Postdoctoral research at the intersection of advanced computing and engineering.
Major in Computer Engineering with a minor in Computer Science.
Collaboration
I welcome research discussions with academics, students, and industrial collaborators working on AI security, distributed inference, efficient computing, and hardware–software co-design.