Описание
The Indexing layer processes live sports broadcasts to produce AI-generated highlights in real time — and the video-analysis pipeline behind it is yours to build and operate. The core challenge isn't just shipping code; it's managing high-performance compute workloads across 30+ event-driven microservices that must auto-scale under extreme live-event traffic spikes without failing.
Задачи
•Architect and operate low-latency, event-driven microservices that handle live event traffic bursts with zero downtime
•Bridge ML models into production by building resilient deployment pipelines that handle model drift, context limits, and non-deterministic outputs
•Own end-to-end service delivery from initial RFC design through zero-downtime deployment and post-launch telemetry
•Optimize pub/sub event pipelines across Kafka and Azure Service Bus to decrease processing latency and boost data throughput
•Debug distributed performance bottlenecks and scale Kubernetes compute clusters handling concurrent 24/7 media processing
•Collaborate with ML Engineers and Data Scientists to design systems that degrade gracefully during inference failures
Требования
•4+ years scaling distributed microservices and backend systems in production cloud environments
•Strong expertise in C# / .NET for high-throughput, low-latency application development
•Production experience with Kubernetes, relational schema design, SQL query performance tuning, and messaging infrastructure (Kafka, Azure Service Bus)
•Hands-on integration of AI-assisted engineering tools (Cursor, Claude Code, sub-agents) into daily workflow
Будет плюсом
•Operational knowledge of production ML/AI behavior (inference performance, context limits, non-determinism)
•Azure cloud infrastructure experience (networking, IAM, cost/performance optimizations) or exposure to video/media processing pipelines
Условия
•Direct ownership of high-concurrency event pipelines where code changes instantly impact live, broadcast-level media products
•Pragmatic adoption of AI engineering practices — AI coding assistants and sub-agents are integrated into production workflows rather than treated as an afterthought
•Exposure to complex distributed systems engineering, balancing massive scale, live media ingestion, and operational cost efficiency
