Perception and sensor fusion
Camera, IMU and LiDAR into one reliable signal.
Applied AI that holds up on real data.
Computer vision, sensor fusion and LLM systems, built to run in production. Based in Turin.
Expertise
LLM and RAG systems
Assistants that answer from your data, with voice.
Real-time performance
Pipelines that run fast, up to 200 fps.
Deployment
Models that ship and keep running.
Selected work
Multi-model AI assistant
MSC
Motion-capture pipeline at 200 fps
Move.ai
GPU-accelerated vision for RL
BlueDynamic
Computer vision R&D lead
Braendler Engineering, published paper
40 projects delivered
Gfuse, founder and CTO
Voice assistant demo
Ask a question. It searches, then answers.
Scripted demo with made-up data. Not connected to a live model or any company system.
Assistantidle
Pick a question to start.
Code examples
- vioC++GPU-accelerated visual odometry with IMU fusion
- Tangent_PlannerC++Obstacle avoidance for mobile robots on ROS
- feature-extraction-openclC++GPU visual-inertial odometry with OpenCL
- CPP_buffer_templateC++Generic buffer for sensor frames and odometry
- Interval_MapC++Range-to-value map built on std::map
- maritime-logisticsRustShipping-support assistant, built for safety and evaluation
- teraform_awsHCLTerraform setup that launches AWS servers
- KnightPythonFinds every shortest knight route on a chessboard
Have a hard technical problem?
- A model that works in a notebook, not in production
- A pipeline too slow for real time
- An assistant that answers from your documents
- Sensor data that has to become one decision
- Research
- Design
- Build
- Test
- Deploy
- Support
Clients
Excellent machine learning and deep learning engineers — highly recommended.
Great collaboration — they delivered, and the results were beyond expectations.
Quick and with no nonsense consultation about deep learning, Docker, and embedded devices.