About the Role You will own the inference backbone behind QVAC's local AI stack: the C++ systems layer that makes models run fast, reliably, and predictably on real user hardware. The role is centered on engineering quality at runtime level, including startup behavior, memory pressure, throughput/latency balance, and long‑session stability. You will define and evolve the core abstractions that inference features depend on, so new capabilities can be added without sacrificing performance or maintainability. This is a role for someone who enjoys low‑level problem solving, clear technical ownership, and building infrastructure that other teams trust in production. Your work directly enables private, on‑device AI experiences and helps set the technical foundation for QVAC’s next generation of peer‑to‑peer AI products. Responsibilities Work on deploying machine learning models to edge devices using the frameworks: llama.cpp, ggml, onnx Collaborate closely with researchers to assist in coding, training and transitioning models from research to production environments Integrate AI features into existing products, enriching them with the latest advancements in machine learning Ensure that the inference layer is stable, optimised, and ready for integration with the rest of the stack Qualifications Excellent programming skills in C++ (experience in JavaScript is a bonus) Strong experience with Llama.cpp and ggml inference engines, which facilitates the deployment of models to specific GPU architectures Good understanding of deep learning concepts and model architectures Experience with transformers, LLMs, Diffusion models Demonstrated ability to rapidly assimilate new technologies and techniques A degree in Computer Science, AI, Machine Learning, or a related field, complemented by a solid track record in AI R&D #J-18808-Ljbffr