How It Works

Benefits

Contact

FlowSense AI

Learning-Based Traffic Analysis for Real-World Transportation Networks

FlowSense builds graph-based traffic models that learn from agency-specific data to support consistent, reliable infrastructure planning.

Import your data.

Upload traffic counts, travel times, and roadway geometry to initialize
a network-aware model.




AI calibration engine.

Learns network-level traffic behavior from historical data using Graph Neural Networks, reducing manual calibration effort.




Validated outputs.

Scenario-ready outputs designed to support corridor and
network-level planning analysis.




What we do?

FlowSense is developing a learning-based traffic modeling platform built on Graph Neural Networks trained on real-world transportation data.


Traditional traffic modeling workflows are powerful but often require extensive manual calibration and are difficult to carry forward across studies. As a result, valuable institutional knowledge is frequently lost between projects.


FlowSense introduces a learning layer that captures recurring traffic behavior across networks, allowing agencies and engineers to evaluate scenarios with greater consistency and less repetitive effort.


We are not replacing existing simulation tools. We are building foundational models that complement current workflows and improve how traffic behavior is represented over time.


Prototype Dashboard

FlowSense maintains a prototype dashboard to demonstrate how our learning-based traffic models can be accessed and explored in practice.


The dashboard connects to our current model pipeline and visualizes network-level traffic behavior, scenario inputs, and resulting performance metrics using real-world roadway data.


This prototype is intended for technical review and early validation. It represents one implementation of the underlying modeling framework, not a finished product interface.

Built on FlowSense’s internal model APIs and cloud-based training and inference infrastructure.


View Prototype Dashboard→

Our Current Focus

Our current focus is building and validating an initial foundational model using real-world roadway data in collaboration with transportation agencies and domain experts.

Our Team


Our founding team combines deep expertise in transportation engineering, data science, and machine learning. Together, we’re uniting domain knowledge with rigorous machine learning methods to advance how mobility systems are analyzed and planned.

Najmeh Jami, PE
CEO and Founder - Transportation engineer with years of experience in traffic modeling, simulation, and corridor studies, leading projects for state and city agencies. Passionate about leveraging AI for smarter infrastructure decisions.

Roy Forestano, PhD
Founding Research Scientist - Machine learning researcher with deep expertise in graph-structured data, geometric deep learning, and sequence modeling. Interested in developing scalable, accurate, and explainable computational models which integrate well-tested theory with real-world mobility networks.

Pieter Moens, PhD
Founding Machine Learning Engineer - Machine learning engineer with end-to-end expertise in graph data science, scalable architectures and MLOps to bridge research, engineering, and deployment for reliable real-world mobility networks.

Jason Morris
Advisor - Business strategist, notable for building scalable tech platforms and leading cross-functional teams. Expert in product vision, operations, and forming impactful partnerships.