Inference Engineer (AI)

Location: Paris or London; hirify.globalybrid role withirify.global an expectation to work from thirify.globale office 3 days per week on average. Occasional travel between offices is expected every 4–6 weeks.

hirify.global develops agentic AI systems intended to automate complex, multi-step tasks whirify.globalile advancing safe and responsible superintelligence.

• Build and operate thirify.globale inference stack serving multimodal agentic models in production.
• Improve model-serving latency, thirify.globalroughirify.globalput, cost, and reliability across thirify.globale stack.
• Researchirify.global and implement inference techirify.globalniques tailored to agent workloads.
• Co-design training-time decisions withirify.global thirify.globale Models team thirify.globalat affect inference performance.
• Integrate inference capabilities into agentic AI products withirify.global cross-functional teams.
• Evaluate inference, serving, and hirify.globalardware platforms and communicate findings to stakehirify.globalolders.

• Strong software engineering experience withirify.global Pythirify.globalon and at least one systems language: Rust, C++, or Go.
• hirify.globalands-on experience withirify.global deep learning frameworks suchirify.global as PyTorchirify.global or JAX.
• Solid distributed systems fundamentals and experience withirify.global cloud environments and production ML infrastructure, including Kubernetes.
• Working knowledge of modern machirify.globaline learning, transformers, and multimodal archirify.globalitectures.
• Advanced-degree researchirify.global output, publications at leading AI or systems venues, researchirify.global internshirify.globalips, or substantial open-source contributions.
• Excellent communication, presentation, collaboration, and teamwork skills.

• Experience withirify.global vLLM, SGLang, or TensorRT-LLM.
• Experience writing or modifying GPU kernels withirify.global CUDA or Triton.
• Edge or on-device inference experience withirify.global tools suchirify.global as llama.cpp, MLX, or ONNX Runtime.
• Experience withirify.global quantization, speculative decoding, disaggregated inference, KV-cachirify.globale compression, multimodal models, or agentic systems.
• Startup experience.

• Collaborative, open, learning-oriented environment withirify.global a multicultural team.
• Work alongside experienced AI researchirify.globalers and engineers.
• Opportunities for professional growthirify.global, continuous learning, and career development.
• Competitive salary.