Senior AI Inference Engineer - Large Language Models (f/m/d)

You will join our product team in a position that sits at the intersection of artificial intelligence research and real-world solutions. We foster a highly collaborative work culture where you can expect to work closely with your teammates and have a high level of communication between teams through methodologies such as pair or mob programming. Your responsibilities: Model Inference: Focus on inference optimization to ensure rapid response times and efficient resource utilization during real-time model interactions. Hardware Optimization: Run models on various hardware platforms, from high-performance GPUs to edge devices, ensuring optimal compatibility and performance. Experimentation and Testing: Regularly run experiments, analyze outcomes, and refine the strategies to achieve peak performance in varying deployment scenarios. Staying up to date with the current literature on MLSys Your profile: You care about making something people want. You want to ship something that will bring value to our users. You want to deliver AI solutions end-to-end and not finish building a prototype. Bachelor's degree or higher in computer science or a related field. You understand how multimodal transformers work. You understand the characteristics of LLM inference (KV caching, flash attention, and model parallelization). You have hands-on experience with large language models or other complex AI architectures. You have experience in system design and optimization, particularly within AI or deep learning contexts. You are proficient in Python and have deep understanding of deep learning frameworks such as PyTorch. A deep understanding of the challenges associated with scaling AI models for large user bases. Nice if you have: Previous experience in a high-growth tech environment or a role focused on scaling AI solutions. Expertise with CUDA and Triton programming and GPU optimization for neural network inference. Experience with Rust. Experience in adapting AI models to suit a range of hardware, including different accelerators. Experience in model quantization, pruning, and other neural network optimization methodologies. A track record of contributions to open-source projects (please provide links). Some Twitter presence discussing ML Sys topics.