{"ref":"content:devtoberfest-2026-ai21-multiple-models","page":"/content/devtoberfest-2026-ai21-multiple-models","json":"/content/devtoberfest-2026-ai21-multiple-models.json","generated":"2026-10-26","drafted_with_ai":true,"entity":{"id":"devtoberfest-2026-ai21-multiple-models","content_type":"talk","title":"🟣 Enable Your AI application to Run on Multiple Models to Ensure Resilience and Compliance","url":"https://www.youtube.com/watch?v=MxwM_5iViIw","published":"2026-10-01","published_on":"source:youtube","author":"SAP Developers","summary":"Devtoberfest 2026 session AI-21. A cross product owner on the [SAP AI Core](artifact:btp-ai-core) team gives the reasons an AI application ends up on more than one model — a model down at the hyperscaler, a data centre that is not allowed to run it, a version being deprecated — and runs a joke app that answers from Gemini 2.5 Flash in EU10 and from Claude 4.5 Haiku in Asia-Pacific. The walk through AI Core's parts is where the edges hold: prompts kept as templates in the prompt registry, imported through the launchpad and the APIs or synced as YAML from a git repository ([Prompt management](capability:prompt-management)); the evaluation service scoring one prompt across up to four models per run on built-in or model-as-judge metrics, and the prompt optimizer generating variants against one metric ([Agent evals](capability:agent-evals)); and the orchestration service harmonising the providers behind one message format with a priority-sorted fallback list ([LLM integration](capability:llm-integration)). The session also states where a metric has to be designed rather than picked: an exact-JSON match scores an email generator at zero, so the joke app's custom metric judges topic and family-appropriateness instead.","short":"Running an AI application on multiple models with SAP AI Core","explains":["artifact:btp-ai-core",{"ref":"capability:prompt-management","applies_to":{"ecosystems":["btp"]}},{"ref":"capability:agent-evals","applies_to":{"ecosystems":["btp"]}},{"ref":"capability:llm-integration","applies_to":{"ecosystems":["btp"]}}],"sources":[{"url":"https://www.youtube.com/watch?v=MxwM_5iViIw","title":"Enable Your AI application to Run on Multiple Models to Ensure Resilience and Compliance — Devtoberfest 2026, automatic transcript","checked":"2026-10-19","quote":"And now we have a fallback mechanism in this orchestration service that essentially you configure a list of models and prompts and filters."},{"url":"https://www.youtube.com/watch?v=MxwM_5iViIw","title":"Enable Your AI application to Run on Multiple Models to Ensure Resilience and Compliance — Devtoberfest 2026, automatic transcript","checked":"2026-10-19","quote":"we have the orchestration service that harmonizes the model languages and APIs into one API."},{"url":"https://www.youtube.com/watch?v=MxwM_5iViIw","title":"Enable Your AI application to Run on Multiple Models to Ensure Resilience and Compliance — Devtoberfest 2026, automatic transcript","checked":"2026-10-19","quote":"we have this prompt registry service um that has two ways of um ingesting templates."},{"url":"https://www.youtube.com/watch?v=MxwM_5iViIw","title":"Enable Your AI application to Run on Multiple Models to Ensure Resilience and Compliance — Devtoberfest 2026, automatic transcript","checked":"2026-10-19","quote":"here we have the evaluation service um within AI core that can take your prompt and for different models um can run evaluations on input data and use evaluation metrics that we provide or that you um configure to give you a use case specific model performance."},{"url":"https://www.youtube.com/watch?v=MxwM_5iViIw","title":"Enable Your AI application to Run on Multiple Models to Ensure Resilience and Compliance — Devtoberfest 2026, automatic transcript","checked":"2026-10-19","quote":"You can use up to four for each run."},{"url":"https://www.youtube.com/watch?v=MxwM_5iViIw","title":"Enable Your AI application to Run on Multiple Models to Ensure Resilience and Compliance — Devtoberfest 2026, automatic transcript","checked":"2026-10-19","quote":"We do have a prompt optimization service."}],"provenance":{"discovered_via":[{"source":"source:youtube","channel":"devtoberfest-2026"}],"first_seen":"2026-10-19"},"drafted_with_ai":true},"readings":{"liveness":[{"label":"Checked against the source","tier":"unknown","evidence":"nobody has checked it"},{"label":"Source reachable","tier":"unknown","evidence":"not crawled"},{"label":"Age","tier":"good","evidence":"published this month"}],"compatibility":[{"label":"Version","tier":"unknown","evidence":"no version declared"},{"label":"Moved since?","tier":"good","evidence":"no recorded change after 2026-10-01"},{"label":"Findings","tier":"good","evidence":"none flagged"}],"reach":[{"value":null,"unit":"views"},{"value":null,"unit":"comments"},{"value":null,"unit":"likes"}]},"signals":null,"series":null,"edges":{"authors":[]}}