Venturi Systems — AI Traffic Optimisation
Founding a company, building the simulation platform, and selling it to city governments — all at once.
City traffic signals were tuned by hand. There was no scalable way to simulate an intersection network and optimise it for real-world safety and flow — and no off-the-shelf path from a research idea to something a municipality would deploy.
- 01
Co-founded and led the company from concept to production deployment, owning business development, product strategy and technical architecture simultaneously.
- 02
Designed a distributed simulation platform that processed 3M+ data points across 140 simulated years of traffic on Google Cloud 96-core instances with Python multithreading.
- 03
Applied ML to accelerate signal optimisation (Green Wave), with matrix-based safety intergreen calculations so the optimisation never traded safety for flow.
- 04
Stood up the delivery pipeline — dockerised service management and CI/CD — so research could ship as deployable infrastructure.
- 05
Owned the commercial side too: secured two business-to-government contracts across 30+ meetings with municipal decision-makers, and managed an external design team plus a small engineering team.
16 intersections optimised and deployed across three municipalities, with partnerships spanning Kyiv, Ivano-Frankivsk and Hong Kong — carrying CTO and CEO responsibility at the same time.
- Python
- GCP
- ML
- Docker
- B2G