Ali Rouhani | OR Scientist • Data Scientist • Data Engineer • Applied AI

Montreal, Canada • ali.rouhani@umontreal.ca

Ali Rouhani

Operations Research | Machine Learning | Data Engineering | Logistics Consulting

Decision scientist and engineer specializing in the intersection of large-scale mathematical optimization, artificial intelligence, and structural data architecture. I develop exact scalable algorithms (adaptive Benders decomposition), engineer robust MLOps data pipelines (XGBoost, Docker, Azure), and build interactive decision-support systems. My work translates complex, high-dimensional uncertainty into actionable, cost-reducing strategies for logistics operations, dispatchers, and corporate stakeholders.

LinkedIn GitHub Scholar

Focus areas

Optimization: Exact solution methods, Benders decomposition, Stochastic programming

Data Engineering: ETL pipelines, PostgreSQL, Cloud Deployments (Azure, Docker)

Applied AI & ML: Ensemble forecasting, classification, scenario clustering

Logistics Consulting: Freight routing, capacity planning, decision-support systems

Keywords

Adaptive Benders LightGBM / XGBoost ETL & SQL Python/C++/Julia LTL / Network Design Stakeholder Communication

About

I bridge the gap between rigorous mathematical theory and production-ready software engineering. Whether I am architecting a centralized SQL database, deploying containerized machine learning models to the cloud, or utilizing C++ solvers to optimize 500,000+ logistics routes under uncertainty, I focus on delivering scalable systems that empower human decision-makers and drive tangible business value.

Impact highlights

Algorithmic Scalability

Overcame monolithic scalability limits by engineering custom exact solution methods (Benders decomposition frameworks), achieving a 60% reduction in computational runtimes for massive-scale logistics networks.

End-to-End Data Pipelines

Architected robust ETL workflows, cloud-based data ingestion (Azure, PostgreSQL), and containerized predictive architectures to reliably transition analytical models into production environments.

Strategic Consulting & Communication

Translated complex mathematical logic and structural data findings into actionable operational insights for capacity planners and dispatchers, presenting strategies globally at technical conferences.

Technical Arsenal

Data Engineering & Cloud

PostgreSQL SQL Docker Azure Linux/Bash ETL Pipelines

Machine Learning & AI

XGBoost & LightGBM Logistic Regression scikit-learn pandas & NumPy Feature Engineering Classification Pipelines

Optimization / OR

Benders Decomposition Dantzig-Wolfe Stochastic Programming MI(N)LP Scenario Clustering

Logistics & Consulting

Network Design LTL/FTL Routing Capacity Planning Inventory Management Decision Support

Programming

Python C++ (C++17/20, STL) Julia pybind11

Solvers & Testing

Gurobi / CPLEX OR-Tools Pytest & GoogleTest GitHub Actions

Professional Experience

Operations Research Scientist & Data Engineer Jan 2021 – Present

CIRRELT, Université de Montréal • Montreal, Canada

Formulated large-scale multi-layer network design models (e.g., coordinating scheduled trains with consolidated freight blocks) to optimize worldwide and regional freight operations under uncertainty.
Engineered a C++ computational engine utilizing Benders decomposition to overcome monolithic scalability limits, successfully solving massive-scale formulations handling over 500,000 arcs.
Implemented an adaptive Benders decomposition algorithm to solve stochastic models under joint capacity and demand uncertainty, reducing median solve times by 60% via parallel evaluation and memory-efficient structures.
Designed a scenario clustering and feature engineering pipeline to compress large stochastic datasets while preserving key decision-relevant probability distributions.
Executed extensive computational experiments on HPC environments, implementing strict unit and integration testing pipelines to guarantee reproducibility across model revisions.
Presented complex supply chain models and predictive frameworks at the JOPT and ISMP conferences, effectively translating mathematical strategies into actionable insights for diverse stakeholders.

Data Analyst & Operations Researcher (M.Sc.) Feb 2018 – Dec 2020

Shahed University

Designed facility location and inventory allocation models to support managers in maintaining resilient service levels under uncertain demand.
Improved the computational performance of an inventory planning model by 50%, accelerating the decision-making process for adaptive stocking and critical supply distributions.
Built data-driven decision-support models for crisis logistics and supply chain resilience, reducing planning runtimes by 40% (with less than 1% deviation from baseline) to manage critical supply allocations during severe network disruptions.

Data Analyst & Software Developer Jun 2017 – Jun 2018

Afrand Sana’t Company

Consolidated and cleaned inconsistent client records, migrating fragmented Excel files into a unified SQL Server database to improve organizational data quality and end-to-end visibility.
Streamlined tracking workflows and optimized SQL queries, accelerating operational reporting and ensuring traceability of client orders for administrative teams.

Selected Engineering & AI Projects

flowbalance - High-Performance Time-Space Engine • GitHub C++, Python, pybind11, Google OR-Tools, Pydantic, Pandas, pytest

Architected an open-source hybrid software package to solve massive-scale dynamic Multi-Commodity Network Flow problems using exact Dantzig-Wolfe decomposition.
Engineered a high-speed C++ pricing engine linked via pybind11, utilizing a cycle-robust Dijkstra priority-queue search to dynamically generate negative reduced-cost paths.
Developed a Python-based Restricted Master Problem (RMP) using Google OR-Tools (GLOP) integrated with a dual-variable pricing filter that significantly reduces subproblem computations.
Designed strict relational integrity and timeline validation schemas using Pydantic to map abstract temporal inventory states, consumption factors, and asset-specific transit costs flawlessly.
Expanded the engine's applicability to offer strategic decision support for empty container repositioning, multimodal freight scheduling, and cross-docking optimization.

Store Sales - Time Series Forecasting (End-to-End MLOps Pipeline) • GitHub Python, LightGBM, XGBoost, PostgreSQL, Docker, Azure, CI/CD

Engineered a high-granularity Multi-Model forecasting system utilizing weighted ensembles of LightGBM and XGBoost, training 33 independent models to isolate family-specific purchasing behaviors.
Implemented Recursive Autoregressive Inference to utilize short-term lags (1–15 days), significantly reducing temporal blindness compared to static bulk-prediction methods.
Constructed advanced feature sets including volatility-weighted moving averages, exponentially weighted averages (EWMA), and deterministic filters for localized economic cycles.
Architected a scalable data engineering pipeline leveraging PostgreSQL and Azure Cloud to manage high-dimensional transactional datasets, ensuring data integrity for large-scale ingestion.
Containerized the full MLOps stack using Docker on Azure Virtual Machines, establishing a Continuous Integration (CI) framework for automated training and evaluation.
Performance: Achieved a competitive RMSLE of 0.41036, approaching the elite benchmark of 0.37687 in a zero-inflated, complex retail environment.

Predictive Modeling & Classification Pipelines (Data Science) Python, LightGBM, XGBoost, scikit-learn, SQL

Built robust data extraction and preprocessing pipelines for customer behavior prediction (Bank Churn) and demographic survival analysis (Titanic dataset).
Conducted extensive exploratory data analysis and engineered numerical/categorical features to improve model generalization.
Tuned hyperparameters for LightGBM, XGBoost, and Logistic Regression models to maximize predictive accuracy.

Vehicle Routing Research Platform (ML + OR) C++, Python, GoogleTest, GitHub Actions

Developed a modular research platform to evaluate advanced optimization strategies and learning algorithms for the Vehicle Routing Problem (Solomon benchmarks).
Implemented baseline algorithms (greedy heuristics, Dijkstra-based routines) in modern C++ with an extensible architecture strictly designed for seamless metaheuristic and machine learning integration.
Investigated learning-assisted routing: route-cost prediction, instance feature extraction, and heuristic selection to reduce runtime while preserving solution quality.
Integrated rigorous unit testing (GoogleTest) and CI/CD pipelines (GitHub Actions) with memory sanitizers to ensure algorithmic reliability.

Publications & Research

Benders Decomposition for Two-layer Network Design with Capacity Decisions

Industry translation: Scalable network investment and capacity planning; faster planning cycles via decomposition.
Methods: Benders decomposition, Valid inequalities, Matheuristics → C++.

Two-layer Network Design under Capacity Uncertainty

Industry translation: Risk-aware capacity decisions via stress-tested network design across scenarios.
Methods: Benders decomposition, Scenario generation, Branch-and-Cut → C++.

Stochastic Two-layer Network Design under Joint Capacity and Demand Uncertainty

Industry translation: Advanced modeling for two-layer networks accounting for bidirectional uncertainty propagation.
Methods: Adaptive Benders decomposition, Scenario clustering → C++.

Dynamic capacity (re)-allocation in collaborative distribution platforms under disruption uncertainty

Industry translation: Operational re-planning under disruptions; dynamic capacity allocation when conditions change.
Methods: Multi-stage stochastic programming, Dual decomposition, Scenario generation → Julia.

Accelerated Benders Decomposition for Stochastic Liner Shipping Fleet Repositioning

Industry translation: Asset rebalancing under uncertainty to improve cost and service reliability in maritime logistics.
Methods: Two-stage stochastic programming, Benders decomposition, Valid inequalities → Julia.

A Lagrangian Decomposition Algorithm for Robust Green Transportation Location Problem

Publication: International Journal of Engineering TRANSACTIONS A: Basics (2019).
Methods: Lagrangian decomposition.

Education & Awards

Education

Ph.D. in Computer Science (Operations Research), Université de Montréal — 2021 – Present (Expected 2026)
M.Sc. in Industrial Engineering (System Optimization), Shahed University — 2016 – 2019
B.Sc. in Industrial Engineering, University of Qom — 2011 – 2016

Presentations & Awards

Conference Presentations: JOPT, ISMP 2024
UdeM Exemption Scholarship for International Students (2021)
Best Undergraduate Researcher, University of Qom (2016)
3rd Place, Tehran Mathematics Olympiad (2006)