Compiler Engineers for Autonomous Driving and Custom AI Hardware – ITALY
Pubblicato il 03-08-2026 - Cortus in Lecce
ph3Compiler Engineers for Autonomous Driving and Custom AI Hardware – ITALY /h3 pCortus is a fabless semiconductor manufacturing group with a presence in several EU countries. The company designs and sells RISC-V microcontrollers (MCUs) and AI inference chips for automotive and avionics markets. Cortus develops a broad portfolio of proprietary IP and provides a complete software ecosystem, including IDE, compiler, debugger, SDK, and development boards. Over 18 billion devices have been produced using Cortus processors and IP, with around 1.2 billion units shipped annually. Cortus is also one of the dozen Platinum Founding Member of the RISC-V Foundation. Cortus is at the forefront of advanced AI inference chip development. /p pWe are seeking Software Engineers (m/f), with a focus on compiler engineering for AI applications, to join our RD Center of Excellence in Lecce. /p h3bRole Overview /b /h3 pWe are building core technology for real-time AI in autonomous driving. Our focus is not on training larger models, but on making them run efficiently and reliably on specialized hardware, this is fundamentally a systems problem. We are looking for engineers to develop the “compiler” layer that takes trained AI models (ONNX) and maps them onto our custom AI hardware for high-performance inference. /p pYou will work at the intersection of:
/p ulliPerformance optimization /li /ul h3bKey Responsibilities /b /h3 ul liConvert ONNX models (e.g. YOLO-style networks) into optimized execution graphs /li liMap neural network layers onto compute resources under memory, bandwidth, and latency constraints /li liOptimize execution through graph transformations (splitting, fusion, scheduling) /li liManage and optimize data movement across the system /li liBuild tools and infrastructure to ensure robustness and scalability /li liCollaborate closely with architecture and system teams /li /ul h3bProfile /b /h3 ul liStrong understanding of how software executes on hardware /li liAbility to reason about performance trade-offs (compute vs memory vs communication) /li liExperience with compilers, graph-based systems, or complex software stacks /li liGood understanding of neural network structures and execution /li liFluent in English (written and spoken) /li /ul h3bNice to Have /b /h3 ul liExperience with compiler frameworks (e.g., LLVM, MLIR, TVM) /li liExperience optimizing large-scale or low-latency systems /li liFamiliarity with ONNX or deep learning frameworks /li /ul h3bWhy Join Us /b /h3 ul liYou will define how AI models actually run—not how they are trained /li liYour work will have a direct impact on real-world performance /li liYou will work on challenging problems at the intersection of software and real-world hardware constraints /li /ul /p #J-18808-Ljbffr