MIAR develops proprietary technologies for artificial intelligence. Its strategy starts with ZeroLLM and extends through Robots and Stratosfera towards robotics and autonomous aerospace systems, with potential commercial applications.
Connections across disciplinesAbstract representation
Proprietary technologies. Hands-on experimentation. An international outlook.
ResearchExperimentationTechnology transfer
00 / Projects under development
A technological core. Multiple application directions.
ZeroLLM is the proprietary inference engine at the core of MIAR’s strategy, already running multiple models. Robots and Stratosfera develop research directions in robotics and aerospace simulation. The objective is to evolve the software core to support these applications: integration is a development path that requires validation.
ZeroLLM aims to make AI model execution fast and straightforward to adopt. A proprietary engine designed for data centre workloads and clients with customised models, with automatic configuration as a development objective.
Today
Working inference across multiple models.
Planned development
Expand performance validation and develop plug-and-play adoption. Efficiency and reduced computing costs are objectives to measure.
Robots develops proprietary software for robotic control, aiming to reduce latency and connect training, simulation and physical behaviour. The project works with a virtual humanoid with 23 degrees of freedom.
Today
Working in simulation with NVIDIA Isaac Sim on an RTX 4070 SUPER.
Planned development
Validate basic capabilities on a physical robot and expand behaviours through training.
Stratosfera develops models and simulations for a drone intended for reconnaissance and defence. The work connects force calculations, flight behaviour and control to assess the feasibility of a stratospheric platform.
Today
Flight and control simulated at different altitudes with NVIDIA Isaac Sim on an RTX 4070 SUPER.
Planned development
Improve simulation realism and develop an initial 6-metre prototype. Prototyping is a future objective, dependent on progress through the DIANA application process.
MIAR welcomes conversations with investors interested in developing its proprietary technologies. The strategy starts with ZeroLLM and extends towards the application directions of Robots and Stratosfera. If you would like to explore an investment, contact us for an initial discussion about development plans and commercial prospects.
MIAR — Milan Institute for Advanced Research is a business focused on advanced research and proprietary technologies. Its strategy places ZeroLLM at the centre of software development, with a commercial direction in AI inference. Robots and Stratosfera extend research towards robotics and autonomous aerospace systems. Development of a shared technological foundation proceeds through experimentation, integration and verification.
01
Mission
Explore scientific and technological questions that call for contributions from several disciplines, combining theoretical research and experimentation.
02
Vision
Build an independent research direction, open to international exchange and able to evolve with the questions raised by science and society.
03
Method
Formulate hypotheses, design experiments and verify results. Rigour, documentation and critical evaluation guide every stage of research.
Research and business development
ZeroLLM has its own commercial direction for data centres and clients with customised models. Development of the software core also aims to support future applications of Robots and Stratosfera. Technologies may be commercialised through software, technological solutions and licensing, according to validated results and customer requirements.
AI research and development in Italy
MIAR engages with businesses, researchers and partners interested in artificial intelligence development in Italy, while remaining open to international collaboration. Its projects connect AI inference, robotic control and simulation with potential commercial applications.
These areas outline MIAR’s research interests. They are starting points for exploring new problems, building connections and welcoming other disciplines.
01
Artificial intelligence and computational sciences
Computational methods for data analysis, learning and understanding complex systems. Study of their foundations, potential applications and evaluation criteria.
02
Software and information processing technologies
Software architectures, methods and tools for information processing. An interest in reliability, efficiency and integration in scientific and technological contexts.
03
High-performance computing and distributed systems
Methods and architectures for demanding computational problems. Parallelism, resource distribution and system coordination as subjects of study.
04
Modelling, simulation and digital twins
Mathematical and computational representations of phenomena and systems. Simulations and digital twins to formulate hypotheses, explore scenarios and compare models with experimental evidence.
05
Robotics, autonomous and embedded systems
The interaction of computation, perception and control. Study of integrated systems, autonomy and behaviour in physical environments, with a focus on experimental verification.
06
Aerospace technologies and stratospheric platforms
Study of systems and technologies for aerospace applications and stratospheric platforms. Modelling, integration and assessment of operating conditions.
Research extends beyond a list. New questions can open new fields.
03 / Research approach
From the question to verification.
A methodological process connecting understanding, development and evaluation. Its stages inform one another: each piece of evidence may call for hypotheses and models to be revised.
01
Problem formulation
Define the question, context, constraints and criteria for assessing possible answers.
02
Study
Examine available knowledge and relevant methods. Identify hypotheses to investigate.
03
Experimentation
Design experiments, collect evidence and compare hypotheses with data.
04
Prototyping
Translate hypotheses into models, software or experimental systems to explore their feasibility.
05
Validation
Assess results, limitations and reproducibility against explicit criteria and defined conditions.
06
Potential applications
Examine opportunities for use and transfer, based on the results and the requirements of the context.
This process describes a working method, not results already achieved.
04 / Collaboration
Research grows through exchange.
MIAR welcomes dialogue with businesses, researchers, universities, industrial partners and funding bodies, including in international contexts.
The form of collaboration is shaped by the problem, shared objectives and the expertise required.
Formulating shared research questions, combining expertise and defining a common research process.
02
Commissioned research
Investigating specific problems with agreed objectives, scope and evaluation criteria.
03
Experimental development
Exploring technical feasibility through models, prototypes and validation activities.
04
Research programmes
An interest in participating in national, European and international programmes, according to relevant themes and opportunities.
05
Technology transfer
Potential development of research results into software, technological solutions and licences, where appropriate.
05 / Contact
Start a conversation. About research and investment.
We welcome collaboration proposals and expressions of interest from investors. Share the project you are interested in and your objectives for the discussion.
Complete the form to prepare an email to info@miadvancedresearch.com. You can review and send it from your email application. The website does not automatically send or save form data.
