Research & Education

TENEVIA specializes in developing digital solutions for generating environmental data, leveraging artificial intelligence, specifically computer vision for image analysis and numerical modeling. The company offers a comprehensive suite of products including sensors (CamLevel, CamFlow) for measuring water levels and flow, simulators (HydroCore, SnowCore) for hydrological and glacio-nival forecasting, and specialized software (FlowSnap) for image-based gauging. Additionally, TENEVIA provides online supervision services (OS) for centralized data management, advanced cloud processing, and customizable dashboards. These solutions are designed for various applications such as natural risks, renewable energies, natural resources, and pollution monitoring, catering to professionals in environmental management.

The Artificial Intelligence Research and Innovation Center (AIRIC) at Politecnico di Milano is dedicated to advancing AI research and fostering innovation. It serves as a crucial link between academic scientific research and the practical demands of industry and public administration. AIRIC actively engages in collaborative projects, developing innovative AI solutions that address real-world challenges. The center's work spans various aspects of AI, contributing to both theoretical advancements and their application in diverse sectors. This initiative highlights Politecnico di Milano's long-standing commitment to being at the forefront of AI research for over 50 years.

Bright Giant provides an AI-powered platform for the advanced analysis of small molecule MS/MS data, designed to accelerate scientific discovery. The core of its offering is SIRIUS, a comprehensive solution that transforms how researchers handle LC-MS data. Key functionalities include LCMS pre-processing with automatic feature detection, alignment, and adduct detection, alongside quality metrics for prioritization. It excels in molecular formula annotation using isotope patterns and fragmentation data, and confidently identifies molecular structures by searching databases. SIRIUS can also predict compound classes without relying on structural databases and offers substructure annotation, visualizing the direct connection between MS/MS spectra and structure candidates. The platform supports automatic spectral matching in various libraries and provides flexible subscription plans tailored for individual researchers to enterprise-scale providers, including academic solutions.

NNPACK is an acceleration package specifically designed to optimize neural network computations on multi-core CPUs. It focuses on delivering high-performance implementations of convolutional neural network (convnet) layers. The tool is not intended for direct use by machine learning researchers but rather provides low-level performance primitives that are leveraged by leading deep learning frameworks such as PyTorch, Caffe2, MXNet, and Darknet. It supports various platforms including Linux, macOS, Android, and iOS, and offers multiple algorithms for convolutional layers, including Fourier transform, Winograd transform, and implicit matrix-matrix multiplication. Implemented in C99 and Python, NNPACK features multi-threaded SIMD-aware implementations and extensive unit test coverage.

OAT (Online Alignment Toolkit) is a simple yet efficient open-source framework designed for running online LLM alignment algorithms. It features a distributed Actor-Learner-Oracle architecture optimized for high efficiency, utilizing vLLM for accelerated response sampling and DeepSpeed ZeRO for memory efficiency. OAT simplifies the experimental pipeline by providing an online Oracle for preference data labeling and real-time model evaluation. Researchers can simulate various feedback types, including verifiable rewards and LLM-as-a-judge, with flexible deployment options for reward models. Its modular structure facilitates rapid prototyping and experimentation, implementing cutting-edge algorithms like PPO/Dr.GRPO for online RL and Online DPO/SimPO/IPO for preference learning, fostering innovation and fair benchmarking.

CRTA (Regional Center of Excellence for Robotic Technology) is a research and development center based at the Faculty of Mechanical Engineering and Naval Architecture, University of Zagreb. It specializes in advanced robotic applications for industry and medicine, aiming to replace traditional automation and human labor with flexible, versatile, and adaptive robotic systems. The center provides excellent working and educational conditions for scientists, engineering students, and medical personnel, serving as a platform for developing, testing, and refining advanced robotic applications. CRTA actively engages in scientific research, projects, and educational initiatives, including collaborations with industry partners like BMW Group on AI-based models for efficient processes.

verl, short for Volcano Engine Reinforcement Learning for LLMs, is an open-source RL training library designed for large language models. Initiated by ByteDance Seed team and maintained by the verl community, it provides a flexible, efficient, and production-ready framework for post-training. Key features include easy extension of diverse RL algorithms through its hybrid-controller programming model, seamless integration with existing LLM infrastructures like FSDP and Megatron-LM, and flexible device mapping for efficient resource utilization. verl is known for its state-of-the-art throughput and efficient actor model resharding with 3D-HybridEngine, significantly reducing memory redundancy and communication overhead. It supports various RL algorithms such as PPO, GRPO, and DAPO, and is compatible with popular Hugging Face and Modelscope Hub models.

The Human-Centered Artificial Intelligence Lab (HCAIL) is a research group based at Seoul National University. Their work explores the intersection of health informatics, human-computer interaction, decision-making, data visualization, and social computing. HCAIL is dedicated to leveraging AI to support individuals with special needs through innovative design and engineering solutions. The lab conducts various research projects, publishes findings, and contributes to the academic community, fostering advancements in human-centered AI applications.

doubleAI focuses on advancing machine learning to achieve next-level AI, specifically aiming to solve complex challenges in computer science, mathematics, and science. The company's research is dedicated to developing superintelligence and improving reasoning capabilities within AI systems. A key offering is WarpSpeed, a technology designed to surpass expert-written kernels, making NVIDIA's best GPU code even faster. This innovation is geared towards scaling expert intelligence and progressing humanity through advanced AI applications, particularly in scientific and technological domains. The tool emphasizes research and development to push the boundaries of AI performance.

AISCIA Informatics offers an AI-powered platform designed to accelerate the discovery of chemicals and materials through the integration of AI, quantum simulations, and advanced data science. The platform, AISCIA Platform™, provides intelligent, scalable, and sustainable solutions that aim to reduce costs, decrease time to market, and improve product performance. Key features include model training and visualization without coding, high-speed predictions for single or batch processing, and smart optimization capabilities for balancing constraints and achieving breakthrough results. AISCIA delivers enterprise-grade AI solutions for various high-impact use cases, including plastic and polymer optimization with EXTRUDAI OPTIMIZER™, oil and gas process control with PetroAI™, battery development with VoltGenius™, concrete formulation with BuildAI™, and solar panel efficiency with SolarSynapse™. They also offer custom solutions for specific industry challenges.

Flow Forecast (FF) is an open-source deep learning framework built on PyTorch, specifically designed for time series forecasting, classification, and anomaly detection. Originally developed for flood forecasting, it now supports a wide range of applications. The library integrates the latest state-of-the-art models, including various transformer architectures, attention models, GRUs, and ODEs. It emphasizes interpretability with easy-to-understand metrics and offers seamless integration with cloud providers like Google Cloud Platform, along with model serving capabilities. Flow Forecast was a pioneer in offering transformer-based models for time series and aims to be an end-to-end deep learning solution.

Deepen AI offers industry-leading multi-sensor LiDAR annotation and labeling tools and services, specifically designed for autonomous vehicles and robotics. The platform focuses on enhancing the speed and accuracy of data labeling for multi-sensor data, including 2D and 3D images, videos, and LiDAR. Key features include advanced 2D & 3D annotation capabilities, AI-powered point cloud bounding boxes and segmentation, and multi-sensor labeling. Deepen AI also provides robust calibration tools for various sensors like LiDAR, camera, radar, and IMU, ensuring data integrity and precision. Additionally, it offers data annotation services with skilled in-house annotators and custom engineering solutions for unique use cases. The platform emphasizes safety-critical AI, with built-in QA workflows, real-time productivity analytics, and compliance with ISO 27001, GDPR, and SOC 2 standards.

DANNCE.ai provides an AI-powered solution for motor assessment, allowing clinicians to gather clinical-grade motor data from patient phone videos. It is designed to address the care gap in progressive diseases like Parkinson's, enabling continuous monitoring rather than periodic snapshots. Patients can record guided movement exams remotely with a caregiver using any smartphone, and the AI extracts objective motor data such as tremor frequency, gait velocity, and tap rhythm. This HIPAA-compliant tool integrates into existing clinical workflows, offering objective data between visits to inform treatment decisions and detect medication wearing-off patterns. It is peer-reviewed and clinically proven, providing a secure and encrypted way to track motor symptoms.

Balthazar is a SaaS platform designed for hardware R&D, providing an intelligence platform for deep tech labs. It automates experiment tracking, prototype management, and result analysis, replacing traditional spreadsheets with structured, searchable data. The platform ensures knowledge retention, facilitates real-time collaboration, and supports AI-driven research by integrating machine learning into experimental pipelines. Key features include version-controlled data, automated experiment tracking via Python, live dashboards for insights, and ML-powered R&D capabilities. Balthazar aims to help deep tech companies iterate faster, scale efficiently, and build on their collective knowledge.

Scidrones is a deep-tech company focused on visual intelligence for decision-making, particularly in environmental monitoring. It leverages drone technology and artificial intelligence to enhance detection and monitoring processes. A key offering is the Coastal Marine Litter Observatory (CMLO), which combines drones and AI to track, map, and monitor coastal litter, generating easy-to-read density maps. This tool is designed for a wide range of users, including scientists, governments, NGOs, and local authorities, enabling informed decisions for a cleaner environment. Scidrones also provides services for marine litter pollution monitoring, infrastructure mapping, and high-detail 2D and 3D mapping products, transforming remote sensing data into precise, actionable insights.

ENOT.ai is a comprehensive framework designed to optimize and accelerate neural networks, particularly for PyTorch and TensorFlow pipelines. It offers two main solutions: ENOT Lite for quick results, providing 2-8x compression for Intel CPU/Nvidia GPU users, and ENOT Pro for maximum efficiency, delivering 4-20x compression for custom models with deep customization options. The platform focuses on boosting AI efficiency by lowering computing power requirements, enhancing speed, and cutting costs without needing hardware upgrades. It supports edge deployment without accuracy loss and ensures data security with on-premises or cloud storage options. ENOT.ai utilizes a multi-faceted approach to optimization, including Layer Filter Analysis, Depth Assessment, Input Resolution Consideration, and Latency Optimization, all integrated with a Python API for easy setup.

EVO Human Performance provides AI-powered solutions for athlete performance monitoring and custom software development. Their technology transforms IMU sensor data and other measurements into actionable insights, enabling teams and athletes to optimize performance and prevent injuries. EVO offers personalized, data-driven insights for athletes, fitness enthusiasts, wellness experts, and sports organizations to make smarter, safer decisions. They also specialize in building digital solutions, including websites and web applications, utilizing the latest technologies. Their flagship product, Artemys, helps users with drills, fatigue monitoring, and smart warm-ups, providing data-driven insights into movement asymmetry and recovery needs.

Anto Biosciences is a frontier biology AI lab focused on making the gut microbiome computable for drug development. They develop novel sparsification and tokenization methods to process vast amounts of microbiome data, building foundation models for microbial communities. The company aims to translate these discoveries into frontier research and clinical outcomes, addressing the microbiome-driven causes of drug response and failures. Their Darwin model series predicts drug toxicity and efficacy, optimizing molecules for broader efficacy by understanding how the gut microbiome controls drug response. Anto Biosciences tackles the challenge of traditional methods failing to separate signal from noise in complex microbiome data, enabling causal modeling of microbial ecosystems and forecasting ecosystem dynamics.

Howest AI Lab is dedicated to bridging the gap between cutting-edge AI research and practical industry applications. The lab actively fosters collaborations among experts across diverse subgroups, ensuring a multidisciplinary approach to innovation. Its research initiatives span critical areas including artificial intelligence, big data analytics, energy transition, and cybersecurity. By focusing on these key domains, Howest AI Lab aims to maximize impact through innovative and future-oriented research, providing valuable insights and solutions for businesses and organizations seeking to leverage advanced technologies.

AAI Innovations GmbH, formerly Automotive Artificial Intelligence (AAI) GmbH, offers TÜV-certified tools designed for ADAS and automated driving. Their product suite includes RepliMap for creating and editing ASAM OpenDRIVE-compliant road networks and 3D scenes for simulation, CORA (Compliance & Regulatory Assistant) which translates complex automotive regulations into actionable intelligence, and SGAF (Safety Guidance and Analytics Framework) for building and analyzing safety cases. The company is currently transitioning to AAI Innovations GmbH, expanding its focus beyond automotive to broader data- and AI-driven solutions, while maintaining its commitment to trust in autonomous technology.

DINO-X-API provides examples for using DINO-X, a unified vision model hosted on DeepDataSpace, designed for open-world object detection and understanding. It offers state-of-the-art performance in open-set detection, including significant improvements in recognizing long-tailed objects. The model accepts text, visual, and customized prompts, generating representations like bounding boxes, segmentation masks, pose keypoints, and object captions. DINO-X supports practical tasks such as Open-Set Object Detection and Segmentation, Phrase Grounding, Visual-Prompt Counting, Pose Estimation, and Region Captioning. It also features a universal object prompt for Prompt-Free Anything Detection and Recognition, and seamless integration with AI tools like Cursor and Claude via DINO-X MCP Server.

GENESIS DATA SOLUTIONS offers AI-supported materials integrity management services, focusing on corrosion management and microbiologically induced corrosion (MIC) management. The company leverages artificial intelligence and expert consultancy to enhance efficiency, accuracy, and predictive capabilities, thereby mitigating integrity risks in critical infrastructure. Their solutions are tailored for industries such as upstream oil & gas, where they provide AI data-driven asset integrity management, and wastewater and sanitary sewer systems, with a focus on concrete microbiologically induced corrosion (C-MIC) management. They also engage in applied R&D for AI/ML tools in corrosion assessment and eco-friendly materials/chemicals for corrosion mitigation.

NXAI is at the forefront of AI technology, specializing in the development of advanced AI models built upon its innovative xLSTM architecture. The company is dedicated to pushing the boundaries of AI, particularly for industrial applications, with a strong emphasis on performance, efficiency, and scalability. NXAI's mission is to deliver powerful, energy-efficient intelligence to European companies, facilitating automation, enhancing decision-making processes, and driving significant technological advancements. Their models are designed for real-world deployment, offering robustness and reliability, and are capable of being deployed in cloud environments or on the edge. NXAI collaborates with leading research institutions, including JKU Linz, and is co-founded by Prof. Dr. Sepp Hochreiter, a pioneer in deep learning.

Axiom Bio offers a cutting-edge solution for drug discovery by building a translational intelligence layer that connects experimental data to human outcomes. The platform leverages AI to accurately predict and understand drug toxicity, aiming to reduce unexpected drug failures. Axiom utilizes a rich dataset of thousands of clinical-stage molecules, profiled across proprietary experiments, to link experimental data with human clinical outcomes. This enables comprehensive clinical risk assessments grounded in AI-powered mechanistic reasoning, revealing a drug’s full risk profile and uncovering mechanisms driving toxicity. The tool helps scientists make better-informed decisions earlier in development, ultimately leading to safer medicines.