Is the Swedish Tax Agency ready for the age of AI?
summary
Swedish Tax Agency already runs 10–15 ML-backed services in production and is building a Kubernetes-based MLOps path: model serving with Seldon Core operator on OpenShift, model images via S2I, Prometheus/Grafana metrics, and separate GPU training clusters. Live systems include a vendor chatbot, speech‑to‑text and translation pipelines, and an in‑house handwritten text recognition model for modern Swedish; platform work is partial but the demos and deployments are working end‑to‑end.
AI-written summary · about
chapters (15)
- 0:00Intro
- 2:10Why is AI important
- 8:47AI is our first choice
- 12:17Applications of AI
- 15:17The Platform
- 22:47No vendor lockin
- 28:17Demo
- 29:47Installation
- 37:07Source to Image
- 38:37Deployments
- 40:37Complex Graphs
- 41:27Java Client
- 43:37Route
- 45:17Next steps
- 46:57Conclusion
related talks
40mInnovation within Bounds
27mEnterprise AI at Scale
35mCase Study: AI-as-a-Service on Kubernetes at Scale and In Production
52mA MLOps Platform for All Data Personas
35mAI and Advanced Analytics applied at
01h 03mAI Innovation for a Smarter Financial Ecosystem
27mFrom Chatbots to Agents: Orchestrating Agentic AI on Existing Enterprise Kubernetes Infrastructure.
15mScaling AI-Driven Financial Services: Ethical ML at Scale