About Me

Who Am I?

Hi, I'm Fabrizio Romanelli!

I am a Technical Leader, Research Engineer, and PhD in Robotics and AI. For the past 20 years, I’ve been living at the intersection of industrial robotics, physics-based simulation, and cutting-edge AI, turning complex electromechanical and sensor challenges into scalable, production-grade software architectures.

Throughout my career across industry and research institutions, I’ve worn a few different hats:

  • The AI & Robotics Specialist: I design deep learning pipelines (like YOLOv8), pioneer the use of World Models to close the sim-to-real gap, and drive the technical vision for next-generation industrial manipulators and autonomous systems (using ROS/ROS2, C++, and Python).
  • The Tech Leader & Manager: I have a proven track record of guiding project visions, managing risks, and coordinating cross-functional teams of more than 10 people, establishing robust CI/CD and cloud infrastructures (AWS) to deliver resilient solutions.
  • The Innovator: I love architecting enterprise solutions that leverage Generative AI and LLMs for advanced project insights and automated analysis.

However, this is just one side of the coin! 🪙

When I’m not bridging the gap between simulation and reality or mentoring a team, you’ll probably find me out exploring the world. I am deeply passionate about traveling, trekking, climbing, and sailing. I love capturing moments through my camera lens, getting lost in a good book, or making some noise playing both the guitar and keyboards. And honestly? I probably forgot something, as I’m always looking to add new activities and interests to the list!

Robotics

Deep learning

Software

Autonomous systems

What I do?

Some of my expertise...

Robotics

Experience on 6+ degrees of freedom industrial manipulators (Comau, ABB, Kuka, Universal Robots), mobile robots (Nomad XR4000 et similia), quadruped robots (HyQ, HyQ-Real) and UAVs.

Programming

Wide and proven experience on C/C++ programming for embedded systems, both real-time and non real-time. Programming with Python, bash shell scripting, Matlab and Fortran.

Deep learning

Strong experience in Deep Neural Network architectures and LLMs, design and training for deep generation models to create realistic synthetic sensor data.

System Software Architecture

Strong experience in system software architecture design and development both for robot systems and applications.

Leadership

Proven experience in leading people in technical teams, from small teams of 3 people to medium teams of 10+ people. Focused on guiding the project technical vision, analyzing risks and requirements with clear and strong communication skills.

Real-time Systems

Wide experience on real-time operating systems: industrial-grade experience on VxWorks and real-time Linux (both RTAI and Xenomai) with low level advanced kernel setup.

Years old
Publications
Languages
Patents
My Specialty

My Skills

I love computers (and game consoles) since when I was a kid. This passion guided me towards the search for ever-increasing stimuli in computer science and technology. The passion for robots came during the second year of the University, where I asked myself: why should I learn programming if I cannot make things alive?
Here follows a short list of my main competences.

System software architecture

C

Real-time OS

C++

Linux

Matlab

Deep learning

Python

CSS

FORTRAN

Industrial communication protocols

Shell scripting

Perl

ROS/ROS2

Javascript

Tensorflow

Neural Networks

Simulink

Education

Education

Research on robotic perception, sensor fusion techniques, SLAM, Visual SLAM and visual odometry using both classical and deep learning techniques. Design of ROS 2 software architectures for autonomous systems. The research also addresses the problem of integrating information coming from different sensors in order to cope with redundancy, heterogeneity and to the faults in state estimation (e.g. arising when a sensor malfunctions or degrades its quality). Design and development of a generic framework for resilient robot perception in degraded environments. Support on the Hyrman project for the remote handling design and simulation in Tokamak fusion reactor. Support on the Electron Cyclotron plant Control System for the architecture definition and conceptual design.

I continued my studies in robotics and automation, specializing in control systems at the University of Tor Vergata in Rome. For my master thesis I've developed a system to make a mobile robot able to localize and map in unknown environments. I've graduated with full marks cum laude.

I studied computer science engineering at the University of Tor Vergata in Rome. Here I focused my studies in basic informatics, automation and robotics. For my degree thesis I developed a vision system for mobile robots.

I attended the Liceo Scientifico P. Ruffini scientific high school in Viterbo, focusing on maths, physics and biology. I also had a strong interest in Latin, philosophy and Italian literature which I studied during my courses.

Experience

Work Experience

Senior AI & Robotics Engineer May 2024 - present

AlmavivA S.p.A.

Generative AI & LLM Production Deployment: Architected, developed, and deployed enterprise-grade LLM solutions leveraging vLLM and PyTorch on high-performance NVIDIA GPU-powered server clusters, streamlining automated project analysis and technical insights. Inference Optimization & Performance Tuning: Spearheaded infrastructure-level optimization, quantization, and performance tuning for large-scale LLMs to maximize throughput and minimize latency in production. Advanced Computer Vision: Designed and implemented robust, industry-grade deep learning models (YOLOv8) for visual object detection and resilient perception within highly cluttered and non-deterministic environments. DevOps & MLOps Infrastructure: Established and managed robust Azure CI/CD pipelines, automating testing, model validation, and deployment workflows for scalable software delivery. Technical Leadership & Agile Practices: Drove engineering excellence using Agile methodologies, mentoring cross-functional team members on modern software versioning, Object-Oriented Programming (OOP) design patterns, and code architecture best practices.

Researcher Dec 2022 - May 2024

Consiglio Nazionale delle Ricerche - CNR

Mission-Critical Control Architecture: Engineered the high-level control architecture and real-time remote monitoring systems for the Electron Cyclotron Resonance Heating (ECRH) unit within the international DTT (Divertor Tokamak Test) nuclear fusion facility. System Design & Power Management: Performed rigorous technical analysis to design automated management systems for high-power gyrotrons and specialized, high-voltage power supplies, ensuring deterministic execution and operational safety. Requirements & Risk Analysis: Acted as a technical vision leader, translating complex thermonuclear fusion physics constraints into concrete software engineering requirements and robust system specifications. Cross-Functional Collaboration: Partnered closely with international research bodies and multidisciplinary engineering teams to drive the integration of distributed control systems for next-generation energy research.

Business Development Manager Jun 2021 - Dec 2022

Agrivol

Autonomous Platform Architecture: Spearheaded the definition of complex technical requirements and core system architectures for automated, field-deployed agricultural robotic platforms. Advanced Deep Learning & Perception: Engineered production-ready, industry-grade software solutions utilizing advanced deep learning models for high-fidelity semantic image segmentation, resilient outdoor navigation, and multi-modal sensor data generation. Strategic Technology Partnership: Served as the primary technical interface and drove the integration of advanced mobile robotics platforms as an official partner with Clearpath Robotics for the PartnerBot development program. Business Strategy & Funding Acquisition: Authored, negotiated, and secured multi-million-euro R&D projects through European and national rural development programs (PSR), directly managing the commercial deployment and go-to-market strategy of the developed robotic solutions.

Training Specialist on industrial robotics Mar 2020 - Present

IN.SI.

Advanced Engineering Training: Designed and delivered high-level technical training modules covering industrial robot control systems, rigid-body kinematics, multi-body dynamics, and complex trajectory planning. Industrial Systems Interfacing: Instructed professional engineers on low-level industrial programming and hardware interfacing protocols between robotic manipulators and PLCs via standard industrial fieldbuses. Interactive Kinematics Simulation: Developed and deployed advanced interactive simulation environments to model, test, and program complex automation workflows for Universal Robots and ABB platforms. Strategic Educational Leadership: Served as the lead subject matter expert for the high-profile automation training framework CO.P.E.R.N.I.CO (COstruzione di imPianti intEgRati iNtelligentI e COnnessi), focused on the deployment of integrated, intelligent, and connected industrial plants.

Senior Cloud Engineer Mar 2019 - Jan 2021

Applaudart

Distributed Systems Leadership: Led the core backend engineering team in architecting and deploying highly scalable, distributed cloud computing infrastructure using Amazon Web Services (AWS). Serverless & Data Architecture: Spearheaded the integration of complex enterprise-class systems utilizing a diverse AWS stack, optimizing high-throughput data pipelines through Lambda, DynamoDB, ElasticSearch, AppSync, and EC2 computing instances. Automated MLOps / DevOps: Designed and maintained production-grade CI/CD automation pipelines, minimizing system downtime and ensuring seamless deployment workflows for highly diverse application services. API & System Integration: Engineered resilient, high-performance REST and GraphQL APIs to support real-time data communication across decoupled, enterprise-level application layers.

Robotics Specialist Jan 2020 - May 2020

University of Rome Tor Vergata

Distributed ROS 2 Architecture: Architected and deployed a low-latency, distributed ROS 2 software framework optimized for real-time perception and autonomy in mobile robotic systems. High-Performance VSLAM Integration: Conducted rigorous state-of-the-art research on Visual SLAM and deep learning techniques, successfully integrating the ORBSLAM2 algorithm into a hardware-in-the-loop (HIL) testing environment. C++ Driver Engineering: Developed highly optimized, low-level C++ wrappers for RGB-D camera drivers to maximize raw sensor throughput and ensure deterministic data streaming. Computational Benchmarking & Optimization: Executed extensive functional testing and algorithmic profiling, establishing standardized performance and computational benchmarks across monocular, stereo, and RGB-D sensing arrays in complex indoor and outdoor environments.

Senior Software Engineer Dec 2017 - Feb 2019

Istituto Italiano di Tecnologia

Technical Leadership & Mentoring: Served as a core member of the Advanced Robotics Research Line, managing the software engineering team and supervising postgraduate students' research and development tasks. Real-Time Control Frameworks: Engineered and maintained robust, low-latency, real-time robot control frameworks and high-throughput communication systems for complex robotic platforms. Legged Robotics Architecture: Designed modular, highly resilient software architectures for legged robots, implementing advanced locomotion, manipulation, and interactive GUI layers. Automated CI/CD & Testing: Managed and optimized advanced CI/CD regression testing pipelines on GitLab; successfully integrated open-source packages and standard ROS ecosystems directly into physical robot hardware-in-the-loop (HIL) environments. Experimental Validation: Led systematic test evaluations and analytical report profiling to continuously validate algorithmic performance and mechanical safety under dynamic operational conditions.

Software Development Manager Mar 2016 - Oct 2017

Comau

Next-Generation Platform Architecture: Directed the technical software design, roadmap definition, and full-stack development of Comau’s next-generation industrial robotics platform across PC, mobile, and embedded real-time controller environments. Cross-Functional Technical Vision: Headed the research and architectural design of multi-platform software ecosystems, ensuring seamless interoperability between high-level user interfaces and low-level robot control systems. Engineering Leadership & People Management: Managed and mentored cross-functional teams of software engineers, fostering talent development, establishing agile engineering best practices, and aligning technical skills with strategic product delivery goals. Resource & Target Optimization: Accountable for project resource allocation, risk mitigation, and timeline management, consistently leading engineering teams toward meeting highly complex technical objectives on time.

Software Engineer Specialist Jan 2009 - Feb 2016

Comau

International Requirements Architecture & Leadership: Led the technical requirements gathering and system specification phases for the brand-new NGRS (Next Generation Robotic Software) platform, successfully coordinating technical inputs across 6 cross-functional international engineering teams. Global Supplier Management: Spearheaded the technical design roadmap and development framework for the new software platform, managing the engineering delivery and integration in collaboration with an offshore Indian supplier. Real-Time Motion Control Systems: Designed, programmed, and analyzed advanced motion control algorithms, kinematic solvers, and trajectory planners for Comau industrial manipulators using C/C++ on Real-Time Operating Systems (RTOS / C4G/C5G control systems). EU Research & Open-Source Infrastructure: Contributed as a core technical expert to high-profile European Union research projects (ROBOFOOT, SMErobotics, X-Act); developed the C4G/C5G Open and Open Realistic Robot Library to support advanced motion planning on open Linux platforms. Algorithmic Optimization (eMotion Project): Directed the internal eMotion project as Technical Head, directly optimizing multi-body dynamic models to maximize robot velocity profile smoothness while significantly reducing mechanical stress and cycle times. Advanced Automation Features & Customer Support: Engineered and deployed specialized industrial features, providing expert-level technical support for complex Conveyor Tracking and real-time collision avoidance within multi-robot Interference Regions.

Robotics Engineer Apr 2007 - Dec 2008

Comau

Innovative Product R&D: Designed and developed core software solutions for innovative robotics products, including the Manual Guidance Device and real-time collision avoidance frameworks within multi-robot Interference Regions. System Test & Validation: Engineered and executed comprehensive testing protocols to validate safety-critical functions, kinematic algorithms, and operational performance across core industrial robot platforms. EU Project Governance: Served as an active member of the Project Management Committee for the European Union-funded ROBOFOOT project, driving the integration of advanced robotic automation and perception solutions into the shoe production industry. Cross-Functional Collaboration: Partnered with multidisciplinary research and engineering teams to translate advanced automation concepts into production-ready software components.

Control Software Engineer Oct 2006 - Mar 2007

ICAPGroup

Automotive System Integration: Commissioned and integrated highly automated, turn-key manufacturing systems and robotic workcells tailored for strict automotive tier-1 production standards. Multi-Platform Robot Programming: Programmed, calibrated, and optimized complex industrial trajectory profiles for Comau and ABB manipulators within high-speed, synchronized production lines. Industrial Vision Systems: Designed and deployed dedicated computer vision software architectures using specialized processing scripts to handle automated quality control and parts inspection. PLC & Logic Control Engineering: Developed and debugged robust deterministic control logic, fieldbus communications, and production workflows using Allen-Bradley PLC systems. Safety-Critical Systems Design: Programmed and validated fail-safe industrial security devices (Pilz hardware) to ensure full compliance with machinery safety directives and human-robot isolation protocols.

IT Consultant Jun 2006 - Jul 2006

Capgemini

Telecom Core Software Engineering: Engineered and optimized critical backend components in C++ for high-throughput billing and rating engines within international telecom infrastructures. Infrastructure Automation & Security: Developed specialized Perl scripts to automate log file analysis, system monitoring, and security audit compliance within Sun Solaris enterprise environments. Enterprise Operations Integration: Integrated and configured HP OpenView Operations (HP OVO) solutions to ensure proactive fault management, real-time alerting, and system stability. International Project Delivery: Collaborated within a distributed, cross-functional team for the high-profile Vodafone Greece project, adhering to strict enterprise delivery timelines and performance benchmarks.

Photography

Portfolio

August 2017 - Namibia
August 2017 | Namibia
August 2015 - Japan
August 2015 | Japan
September 2015 - Japan
September 2015 | Japan