Giovanni Claudio

Giovanni Claudio

Sr. Principal Software Engineer @ Medtronic

Giovanni Claudio

Senior Principal Software Engineer at Medtronic, working on a high-performance video and AI pipeline that provides real-time guidance during surgical procedures, for both robotic-assisted and conventional laparoscopic surgery.

Born in Genoa, the land of pesto, focaccia, and Genoa CFC, a football club I keep supporting against better judgement. My surname is Claudio (yes, also a first name, no, I can't explain it). When AI finally replaces us software engineers, I've got two backup plans ready: gelato sommelier or cat photographer.

Download CV See my work →

Currently

Full timeline →

Sr. Principal Software Engineer, AI-Video Technical Lead

Medtronic, Digital Surgery

Jan 2025 - Present London, UK

Leading architecture and delivery of a regulated AI-video platform on embedded GPUs, enabling multiple AI applications to share a single GPU while meeting strict medical isolation and latency requirements.

Impact

  • Defined and owned the software architecture for a regulated AI-video platform, enabling multiple AI applications to share a single GPU with strict medical/non-medical isolation and sub-frame overlay latency for real-time surgical guidance.
  • Built a C++ stack (AI manager, scheduler, TensorRT inference, video/AI recording) with reusable libraries for message-passing middleware, visualization, logging, telemetry and monitoring.
  • Led cross-functional AI-video delivery across 8 teams: ML, Systems, DevOps, Cloud, Quality, Regulatory and UI.
  • Co-located architecture and documentation with code, automating variant-aware regulatory artifact generation in CI for traceable, submission-ready FDA deliverables.
  • Introduced an agentic-AI workspace adopted daily for development, documentation, compliance auditing, FMEA review and regulatory preparation.
  • Mentored 5+ engineers through code pairing and technical reviews, increasing delivery pace and design quality.
Role notes ▸

Stepped into the AI-Video Technical Lead role to own the architecture and delivery of Medtronic's regulated surgical AI-video platform, a system that delivers real-time AI guidance during robotic and laparoscopic surgery, from prototype to release.

The platform is built around three goals: isolation (multiple AI applications sharing a single embedded GPU without interference), segregation (clean medical/non-medical software boundaries for IEC 62304 compliance), and performance (sub-10 ms inference latency on every surgical frame). Achieving this required a modular multi-process design with one responsibility per service, a pub/sub and shared-memory IPC layer for zero-copy video transport, and an AI manager that controls inference end-to-end, preventing parallel overloads, enforcing valid AI combinations, and recovering from failure modes.

Beyond the runtime, the role spans the full delivery stack: unified dev containers for engineers and CI, a modular build system enabling feature-gated releases across product variants, shared platform libraries, hardware-in-the-loop integration testing, and documentation practices that feed directly into FDA submission artefacts. As technical anchor across ML, Systems, Quality, Regulatory, and UI teams, this role owns technical decisions from architecture through regulatory sign-off.

platform-architecture technical-leadership cpp tensorrt cuda gpu-scheduling fda-submissions