cv

Research Assistant

Systems Optimization and Computing Laboratory, University of Houston

Designed scalable algorithms for last-mile logistics problems using column generation, deep learning, and combinatorial optimization.

• Developed Branch-and-Price-and-Cut algorithm for drone-based delivery systems.
• Integrated reinforcement learning for subproblem prioritization in column generation.
• Formulated Chvátal-Gomory cuts as a shortest-path-based subproblem.
• Implemented high-performance dynamic programming in Cython.

Graduate Research Assistant

Transportation and Logistics Lab, SUNY Stony Brook

Worked on transportation network resilience and healthcare logistics models.

• Built forecast models for NYC transit recovery after snowstorms.
• Developed stochastic scheduling models for home healthcare.

Ph.D.

University of Houston

Industrial Engineering

M.Sc.

SUNY Stony Brook

Civil Engineering - Transportation

M.Sc.

Sharif University of Technology

Industrial Engineering

Coordinated Mothership Scheduling and Drone Routing Optimization: A Branch-and-Price-and-Cut Approach

Manuscript in preparation

Exact branch-and-price-and-cut with a bidirectional-labeling ESPPRC pricing subproblem and a hybrid cover / Chvátal-Gomory / conflict inequality.

Resource-Based Time and Cost Prediction in Project Networks: From Statistical Modeling to Graph Neural Networks

arXiv preprint arXiv:2511.15003

GraphSAGE and Temporal Graph Networks over heterogeneous activity-resource graphs, reducing MAE by 23-31% against conventional baselines.

Resilience Analysis of New York City Transportation Network After Snow Storms

Transportation Research Record 2677(1), 694-707

Graph-theoretic resilience indices and recovery forecasts for the NYC network across eight snow events.

Scheduling Diagnostic Testing Kit Deliveries with the Mothership and Drone Routing Problem

Journal of Intelligent & Robotic Systems 105(2), 38

Coordinated truck-and-drone delivery scheduling for diagnostic testing kits during the COVID-19 response.

Drone Delivery Optimization

Developed Branch-and-Price-and-Cut algorithm for a mothership-based drone routing problem with novel cuts and reinforcement learning acceleration.

• Bidirectional label-setting algorithm in Cython
• Chvátal-Gomory cut generation with DRL
• Reduction of pricing subproblem evaluations by 30%

Stock Forecasting with LSTM

Built an LSTM model for daily Apple stock prediction using 8 years of data. Achieved 96% directional accuracy and 35% improvement over baselines.

CIFAR-10 Image Classification

Optimized CNN model hyperparameters using Optuna; achieved 80%+ validation accuracy with PyTorch and GPU pipeline.

English-to-French Translation with T5

Fine-tuned T5-small on the OPUS Books corpus with FP16 training; BLEU improved 61% and validation loss fell from 2.14 to 1.47.