01
Speculative Decoding
Faster and fairer LLM inference across heterogeneous and resource-constrained edge systems.
Sydney, Australia · AI Systems
Quan La
Efficient intelligence for distributed and edge systems.
Ph.D. candidate at the School of Computer Science, University of Sydney, and lecturer at Phenikaa University. I work on speculative decoding, federated learning, and efficient AI systems.
CURRENT FOCUS Speculative decoding for edge inference
Researching practical learning and inference under compute, communication, and resource constraints.
Research themes
Systems-minded AI, from algorithms to the edge.
02
Federated & Edge Learning
Communication-efficient collaborative learning that respects resource and deployment constraints.
03
Distributed AI Systems
Algorithms and optimization methods for dependable learning and inference across networked systems.
Recent signals
Selected work.
2026 INFOCOM GoodSpeed: Optimizing Fair Goodput with Adaptive Speculative Decoding in Distributed Edge Inference 2026 IEEE NL FORCE: Federated Orthogonality-aware Low-Rank Adaptation for Communication-Efficient Fine-Tuning 2026 KBS ALSB: A Problem-Aware Hybrid Optimization Framework for k-Coverage and k-Connectivity
About
Research, teaching, and collaboration.
I am a Ph.D. candidate at the School of Computer Science at the University of Sydney, Australia, and a lecturer at Phenikaa University, Vietnam. My research sits at the intersection of efficient AI systems, distributed computing, and machine learning.
I am particularly interested in speculative decoding, federated learning, and practical AI for resource-constrained edge systems. I have been fortunate to learn from Prof. Nguyen Tran and members of the DUAL Group.
For research collaboration, student enquiries, or talks, please reach out by email.