Skip to content

feat(runner): build algorithms with runtime models - #912

Draft
afourniernv wants to merge 1 commit into
NVIDIA-NeMo:mainfrom
afourniernv:codex/litellm-runtime-models
Draft

afourniernv wants to merge 1 commit into
NVIDIA-NeMo:mainfrom
afourniernv:codex/litellm-runtime-models

Conversation

@afourniernv

@afourniernv afourniernv commented Oct 5, 2026 •

Copy link
Copy Markdown
Contributor

What

Adds AlgorithmSpec::build_with_runtime_models(...), which builds an algorithm and the RuntimeModels it needs from the same route configuration.

Why

Embedded Rust hosts can already build an AlgorithmSpec, but the runtime model-group mapping is private. Without this method, a host has to duplicate Switchyard's algorithm-specific mapping for Stage, classifiers, subagents, and the other route types.

Notes for reviewers

Start with crates/switchyard-runner/src/algorithm.rs. The native deployment loader now uses the same builder instead of maintaining its own runtime-model resolver.

This adds one public method. AlgorithmSpec::build(...) and its behavior remain unchanged.

Related: #922 adds a separate error variant for preserving native host errors. The two PRs can merge independently.

Validated with:

  • cargo fmt --all --check
  • cargo test -p switchyard-runner

Signed-off-by: Alex Fournier <afournier@nvidia.com>
@afourniernv
afourniernv force-pushed the codex/litellm-runtime-models branch from a8dfbe1 to e773924 Compare October 6, 2026 17:37
@afourniernv afourniernv changed the title feat(runner): support embedded Rust hosts feat(runner): build algorithms with runtime models Oct 6, 2026

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant