Senior Staff Scientist
AI & Optimization
I am an Applied Science Leader with over a decade of experience at the intersection of optimization, machine learning, and large-scale systems.
Currently, I am a Senior Staff Scientist and AI Lead at Huawei Vancouver Research Center, where I lead a team of researchers building AI capabilities for the OptVerse Optimization Solver and other Huawei Cloud products. Our production routing system supports 40+ VRP variants, serves 20+ enterprise customers, and generates $5M+ in annual revenue. In 2026, our team won 1st place in the CVRPLIB Best Known Solutions competition, producing 51 new state-of-the-art results for large-scale vehicle routing.
My work focuses on LLMs and multi-agent systems for optimization, including LLM-driven algorithm evolution, natural language to optimization modeling, and agentic systems for operations management. I have published at NeurIPS, AAAI, ACL, EJOR, and other top venues, and hold 7 patents (4 published) in AI-augmented optimization. My work has been recognized with multiple awards inside Huawei, including R&D Outstanding Team Award, Future Star Award, and High Value Patent Award.
Previously, I was a Postdoctoral Fellow at Polytechnique Montréal / CIRRELT under Prof. Michel Gendreau, and earned my Ph.D. from Koç University.
LLMs & Agentic AI for Optimization
NL-to-optimization modeling, tool-augmented agents, LLM-driven code evolution
Multi-Agent Systems
Memory architectures, workflow optimization, inference-time search
Large-Scale Combinatorial Optimization
Vehicle routing, network optimization, schedulin (algorithms deployable in real systems)
Our OptVerse-CityU team won the gold standard benchmark for vehicle routing, producing 51 new Best Known Solutions across 100 ultra-large CVRP instances (1,000–10,000 customers) by combining LLM-driven Evolution of Heuristics (EoH) with a massively parallel AILS-II search framework — an effort I led from the OptVerse side.
SmartAPS is an agentic conversational system that transforms how operations planners interact with Advanced Planning Systems. Using tool-augmented LLMs with retrieval-based API selection, it enables natural language what-if and why-not scenario analyses, reducing consultant dependency from days to hours.
Started in 2022, this project aims to democratize operations research by using AI to automatically convert natural language business problems into mathematical optimization models. We established the first benchmark in the field with NL4OPT (presented at NeurIPS 2022 Competition), developed multi-agent systems, created the first reasoning benchmark in OR modeling (ORQA), and introduced graph-based model evaluation metrics.
VRP-Agent automatically analyzes natural language routing problem descriptions to detect VRP features (capacity, time windows, fleet composition, pickup-delivery, etc.), classify the VRP variant, and recommend an appropriate solver — with confidence scores, detailed reasoning, and an interactive Q&A clarification system.
Proceedings of the AAAI Conference on Artificial Intelligence (2025)
INFOR: Information Systems and Operational Research (2024)
arXiv preprint (2025)
Physical Review E (2024)
Journal of Computational Science (2024)
How I automated ArXiv monitoring into a four-layer AI pipeline that reads papers, detects emerging research fronts, and delivers daily briefings.
Insights from the latest advances in AI-driven optimization presented at NeurIPS 2024.
A deep dive into production scheduling challenges and algorithmic solutions.
Examining the role of equity in disaster management and resource allocation.
Feel free to reach out for collaborations, research opportunities, or just to connect.