On-device intelligence
Perception and decision models optimized to run on embedded hardware, with low latency, no data leaving the device, and no dependence on a network link.
Applied AI · Edge-AI
Intelligence that runs where the data is.
URBANGRIDS builds AI systems that perceive, reason and act on the device itself — close to the sensors and the people who use them, fast and private by design, with the cloud as an option rather than a dependency.
What we do
AI from the cloud to the edge.
Three lines of work share one engineering core: models that turn raw data into decisions, deployed wherever they need to run — on an embedded board, a workstation or in the cloud.
Language & data intelligence
Language models, retrieval and machine-learning pipelines that read, classify and act on documents and data, including automation for regulatory and compliance work.
Autonomous systems
Edge-AI autopilots and control software that let uncrewed vehicles and robots navigate, avoid obstacles and manage their missions on their own.
Technology stack
Built to decide on the device.
Our stack keeps the full loop, sensing to acting, as close to the data as possible, and uses the cloud where it adds value: training, fleet-wide learning and heavy analysis.
- 01
Perception & sensor fusion
Computer vision and multi-sensor fusion that turn cameras, radar and other inputs into one coherent picture of the environment.
- 02
Efficient on-device inference
Models compressed and tuned for embedded processors and accelerators, delivering real-time results within tight power budgets.
- 03
Decision & control
Planning and control that act on what the models see: path planning, obstacle avoidance and adaptive mission management.
- 04
Language models & retrieval
LLM-based assistants and retrieval pipelines grounded in an organization's own documents, with every answer traceable to its source.
- 05
Anomaly detection & learning
Models that flag abnormal behavior in systems and data streams, and improve with every deployment.
- 06
Digital twins & cloud-to-edge ops
Simulation for testing before deployment, and a pipeline that trains in the cloud and ships validated models to the edge.
R&D programs
Research backed by Portugal and the European Union.
We develop our technology in competitively selected programs, co-funded by national and European instruments.
PRR · NextGenerationEU
2025 → 2027
AI Applied to Productivity and Sales
Recovery and Resilience Plan project
Development of algorithms that automate the collection and analysis of regulatory and compliance information, together with an internal AI assistant for client support and triage, strengthening the company's AI team and computing capacity.
- Project
- No. 27564 · Call 03/C05-i14.01/2025 — AI in SMEs
- Investment
- C05-i14.01 — Business Innovation
- Period
- Dec 1, 2025 – Nov 30, 2027
- Eligible cost
- €230,150.08
- EU support
- €172,612.56 (NextGenerationEU)
Supported by Portugal's Recovery and Resilience Plan. Project details · recuperarportugal.gov.pt
The company
A small, focused AI engineering company in Portugal.
URBANGRIDS was founded in 2024 in Portugal. We combine machine learning with software, electronics and systems engineering, and we work with research centers, universities and industrial partners across Europe to take AI from the lab into real-world deployment.
Edge first. Decisions made where the data is created, where latency, privacy and connectivity can't be taken for granted.
Open by design. Modular architectures that take in new models, sensors and data sources without starting over.
Validated, not promised. Simulation, measurable benchmarks and structured testing with independent partners.
2024
Founded in Portugal
100%
International client base
Edge+
Cloud and hybrid deployment
EU
Co-funded research and development
Work with us
Let's put AI where it matters.
Partnerships, pilots, research collaborations or product inquiries — we'd like to hear what you're working on.
- info@urbangrids.dev
- Company
- URBANGRIDS
- Headquarters
- Portugal