Head of AI Engineering (f/m/x)

english communication technology german responsibility training safety quality java ci/cd security incident response governance team leadership strategy python aws kubernetes llm pytorch delivery collaboration caching evaluation learning node.js embeddings apis terraform prometheus grafana code quality infrastructure equipment architecture wellbeing microservices observability gemini personal growth backend services capacity planning messaging budgets stakeholder management impact backend engineer team spirit performance incident management platform growth code review team events distributed systems nestjs iac cost optimization organizational development rag openai startup deployment compensation slas llamaindex hiring ai tools commitment distillation mlops proxies orchestration langchain autoscaling ownership ai engineering health benefits security best practices streaming cost tracking feature delivery talent rbac opentelemetry visa sponsorship fine-tuning data privacy a/b tests faiss pinecone cost reporting top talent cost monitoring slos reproducibility aws bedrock llmops operating rhythms cto jvm vllm roadmap management delivery management secrets management bedrock on-call high-growth scalable inference data services batching platform adoption model registries vector dbs api gateways hierarchy ingestion model lineage schemas fitness benefits measurable outcomes triton on-time delivery postmortems auditability peak performance qdrant runbooks exceptional results quantization tgi safety guardrails fast decision-making rate limits product sense rollback ml engineering user-facing features data curation vector stores event-driven microservices onnx runtime multi-provider routing model lifecycle dynamic routing build vs buy systems at scale low-latency inference technical bar fallbacks generative systems cloud ai prompt pipelines vendor strategy canary tests drift monitoring sub-teams research review llm gateway proxy layers high-performance rag async contracts eventing patterns prompt lifecycle prompt registries offline evals online evals inference throughput ai/ml roadmap compliance practices prompt lineage production ai/ml llm stacks millions of daily api calls gpu serving accelerator serving token budgets red-teaming unified llm gateway standardized observability average latency cost per request sub-team structure prompt registry rag services modern ai european tech champion high-end offices right to work in eu