Trainings, Workshops, Coaching
Topics grow out of the real problems we solve in TCW platform practice — not from a training catalogue. Every format is tailored to your stack. No generic cloud-native slide deck.
Agentic AI & LLM Engineering
Agentic AI in Production
What it takes to run agentic AI systems in production — from someone who does exactly that: RAG pipelines with hybrid retrieval (BM25 + vector), embedding selection, cross-encoder reranking, permission-aware retrieval. Tool calling with JSON Schema and Zod, handling malformed model output. Guardrails: prompt injection defence, PII handling, GDPR-compliant design, EU-hosted inference. Evaluations with golden datasets from real production failures.
- Every pattern from a system handling five-figure daily volumes
- Failure modes and incident handling from real operations
- Cost and latency optimisation: prompt caching, model routing, smaller models where they fit
- Your stack or a production-grade demo environment
LLM Orchestration with Spring AI & Kotlin
Build agent systems in the JVM ecosystem: Spring AI for model access, tool calling and structured outputs; Kotlin for domain modelling and non-deterministic output typing. Orchestration patterns: single agent with tools, multi-agent handover, human-in-the-loop approval steps. Monitoring and tracing across the full request path.
- Spring AI, Kotlin, JSON Schema — no Python required
- Typed agent design: model output you can compile against
- Observability: tracing, eval hooks, production dashboards
Kubernetes
Kubernetes 101
Practice-oriented introduction: pods, deployments, services, config maps and secrets, basic cluster administration. You deploy your own app, expose a service, debug a broken configuration.
- kubectl, kind/k3s, Helm basics
- Live demo: deploy, scale and debug a hello-world app
- Logging with stern, metrics quickstart
- GitOps demo with ArgoCD
Kubernetes Enterprise
Multi-cluster, multi-tenancy, network policies, service mesh, storage classes, backup strategies, air-gap deployments. For teams that don’t just test K8s but operate it in production.
- Cilium, Calico, network policies in practice
- Multi-tenancy patterns (namespaces vs. virtual clusters)
- Backup with Velero, disaster recovery
- Cluster API, GitOps for clusters themselves
CI/CD and Platform Engineering
CI/CD Pipeline Build & Handover
Pipeline design without a supply chain of 600 actions. Container-first CI, GitOps with ArgoCD/Flux, build caching, test pyramid, secret hygiene. On your stack, not on a demo repo.
DevSecOps Automation
Security in the pipeline, not in the spreadsheet: SAST, DAST, secret scanning, image signing, SBOM. Compliance (DSGVO, ISO 27001, BSI C5) as code.
Self-Contained Systems & Microservices
Cutting strategies, deployment topologies, inter-service auth. When SCS, when a microservice, when a monolith is the better answer. With a concrete refactoring plan on your codebase.
Identity & Telemetry
OIDC and Token Exchange in Practice
OIDC as the language of your platform. When authorization code, when client credentials, when token exchange (RFC 8693). Multi- tenancy, service-to-service auth, audit trails. With concrete mappings to Keycloak/Zitadel.
Observability without a SaaS Vendor
Metrics, logs and traces on your own metal: Prometheus, Loki, Tempo, Grafana. Server-side analytics without sending behavioural data to a hyperscaler.
EU-Sovereign Platform Migration
An honest path out of AWS/Azure/GCP — without sliding into the next lock-in. Data flows, compliance checklist, migration sprint plan.