NVIDIA/digital-health-skills
HTML
Captured source
source ↗NVIDIA/digital-health-skills
Description: NVIDIA Digital Health skills repo for agent-guided healthcare AI workflows, including evaluation, data generation, model adaptation, deployment guidance, and developer best practices.
Language: HTML
License: Apache-2.0
Stars: 2
Forks: 0
Open issues: 1
Created: 2026-05-14T18:20:29Z
Pushed: 2026-05-28T21:17:06Z
Default branch: main
Fork: no
Archived: no
README:
NVIDIA Digital Health Examples
Agent skills and worked examples for healthcare AI workflows from NVIDIA Digital Health.
This repository hosts agent skills — markdown-based workflow guides consumable by AI coding assistants (Claude Code, Cursor, GitHub Copilot, and any client that supports the agentskills.io specification).
The first set covers a clinical ASR (automatic speech recognition) evaluation workflow: term curation, synthetic clinical-speech benchmark generation, KER (Keyword Error Rate) / entity-level scoring, and fine-tune guidance.
These skills are docs-only workflow guides. They do not ship companion runtime code; an agent following them will use the user's own environment and any required NVIDIA APIs (NIMs at build.nvidia.com, NeMo containers, etc.).
Skills in this repo
| Skill | Stage | What it guides | |-------|-------|----------------| | [digital-health-clinical-asr-setup](skills/digital-health-clinical-asr-setup/) | 1 | Verify NVIDIA_API_KEY, install Python deps, set up NGC + Docker for the NeMo training container, smoke-test the NVCF stack with Magpie TTS + Parakeet/Nemotron ASR. | | [digital-health-clinical-asr-build](skills/digital-health-clinical-asr-build/) | 2 | Specialty interview, term curation, two-tier IPA tagging, NeMo-format manifest synthesis. Inlines a Magpie TTS NVCF recipe. | | [digital-health-clinical-asr-eval](skills/digital-health-clinical-asr-eval/) | 3 | Transcribe a NeMo manifest via Parakeet/Nemotron ASR, score WER/CER/KER/SER, produce a five-section leaderboard, route via the post-eval decision tree. | | [digital-health-clinical-asr-finetune](skills/digital-health-clinical-asr-finetune/) | 4 | Stock NeMo SFT on Parakeet TDT v2, term-aware train/val split, offline cycle N+1 re-eval, optional Riva NIM deploy. |
Each skill folder contains SKILL.md (the workflow guide), optional references/*.md (deeper detail loaded on demand), and evals/evals.json (trigger / behavior / boundary test cases).
Getting started
Skills are loaded by your agent's skill loader. For Claude Code:
git clone git@github.com:NVIDIA/digital-health-skills.git cd digital-health-skills # Either symlink or copy the four skill folders into your Claude Code skills dir: mkdir -p ~/.claude/skills ln -s "$(pwd)/skills/digital-health-clinical-asr-setup" ~/.claude/skills/ ln -s "$(pwd)/skills/digital-health-clinical-asr-build" ~/.claude/skills/ ln -s "$(pwd)/skills/digital-health-clinical-asr-eval" ~/.claude/skills/ ln -s "$(pwd)/skills/digital-health-clinical-asr-finetune" ~/.claude/skills/
Then invoke them from your agent: /digital-health-clinical-asr-setup, /digital-health-clinical-asr-build, /digital-health-clinical-asr-eval, /digital-health-clinical-asr-finetune.
The skills walk the agent through each stage and hand off to the next. Start with /digital-health-clinical-asr-setup.
Requirements
To use the skills end-to-end you will need:
- An
NVIDIA_API_KEYfrom build.nvidia.com — for hosted Magpie TTS + Parakeet/Nemotron ASR via NVCF (free tier is sufficient). - An
NGC_API_KEYfrom ngc.nvidia.com — only needed for Stage 4 (fine-tune), to pull the NeMo training container. - Python 3.10+ for client-side scripting (the agent installs deps via your skill of choice).
- A CUDA host with the NeMo container (
nvcr.io/nvidia/nemo:25.11.01) for Stage 4. Brev cloud GPUs supported.
Support
- Level: Experimental — the skills are new and evolving; expect rough edges.
- Issues: Use this repository's [Issues](../../issues) for bug reports and feature requests. Do not file security issues there (see
SECURITY.md).
Contributing
This is a reference samples repo maintained by the NVIDIA Digital Health team — external pull requests are not accepted. See [CONTRIBUTING.md](CONTRIBUTING.md) for the full policy and what to do if you found a bug (open an [Issue](../../issues)) or want to adapt the skills for your own use (fork under the Apache 2.0 license).
Security
See [SECURITY.md](SECURITY.md) for the NVIDIA vulnerability disclosure process. Report security issues to psirt@nvidia.com, not via GitHub.
License
Apache License 2.0 — see [LICENSE](LICENSE).
Notability
notability 2.0/10Low stars, routine repo
NVIDIA has a repo signal matching data demand, evals and quality, product and customer.