Public summary
An innovative AI cloud infrastructure company is seeking a Senior Applied Scientist to lead research and production optimization projects focused on efficient large language model (LLM) and vision-language model (VLM) inference. The role involves designing and executing research programs, building prototypes, collaborating with engineers to productionize solutions, and publishing credible technical work. The position requires a PhD in a relevant field, strong coding skills in Python and PyTorch, and deep knowledge of transformer inference and model compression techniques. The company offers a collaborative, international environment with opportunities for career growth and impactful AI projects.
Location and work setup
- Location
- Berlin
- Remote status
- On-site
- German requirement signal
- No German Required Detected
- Detected job language
- English
Responsibilities
Lead focused research projects from hypothesis formation through experimentation, prototyping, and production handoff. Prepare technical reports, blogs, or papers to share credible results. Collaborate with machine learning engineers and platform engineers to ensure research outputs become production-ready components. Invent and productionize methods for quantization, QAT, distillation, speculative decoding, KV-cache reuse and compression, long-context inference, MoE routing, and model/runtime co-optimization. Develop rigorous evaluation methodologies covering multiple performance and quality metrics. Mentor engineers and scientists on experimental design and system trade-offs.
Qualifications
PhD in computer science, machine learning, ML systems, computer architecture, electrical engineering, applied mathematics, or a closely related field. Strong publication record or equivalent research artifacts in ML, ML systems, or efficient inference. Proficiency in Python and PyTorch with ability to rapidly develop experiments and prototypes. Deep understanding of LLMs, VLMs, transformer inference, decoding algorithms, model compression, and production-serving tradeoffs. Strong experimental design and statistical analysis skills. Excellent communication abilities. Experience with relevant tools such as Triton, CUDA, or inference-serving frameworks is a plus.