Accessible Neural Search and Cloud-Nativeness for ML Engineers
summary
Open-source Jina provides DocArray for unified multimodal data, a minimal Python refactor into executor + flow units that spawns containerized microservices with async non‑blocking runtimes, a Gateway that routes over gRPC/HTTP/WebSocket/GraphQL, and automatic replicas and sharding for robustness; demo shows a semantic search built in 18 lines of code and deployment to Jina Cloud with a one-line deploy; integrates vector backends (Elasticsearch, SQLite, AnnLite), Prometheus/Grafana observability, and optional Kubernetes orchestration.
AI-written summary · about
chapters (7)
- 0:00Start
- 8:51Unified Interface for Multimodal Data
- 13:40Production ready?
- 21:50Topology Graph
- 23:10Connection Pool RPC
- 24:02Request Streamer
- 28:16Deployment made easy
related talks
35mCloud-Native MLOps Platform for Python- and ML Engineers
01h 11mTutorial: Beyond Neural Search: Hands-on Building Cross-Modal/Multi-Moda...- 35mCloud-Native MLOps Platform for Python- and ML Engineers
26mAdvancing Multi-Modal Search Capabilities in Search Pipeline
19mScalable Data and AI Solutions with Cloud-Native Architecture
03mMaking real-world distributed deep learning easy with Nauta
27mGen-AI at Scale: Simplifying Orchestration of Healthcare Applications Across Multi-Cluster Environme
39mFrom data science to scalable NLU and vision cloud service.