acnlabs · NeilJo-GY · Aug 11, 2026 · Aug 11, 2026 · Aug 11, 2026 · Aug 11, 2026
diff --git a/skills/persona-seed/SKILL.md b/skills/persona-seed/SKILL.md
@@ -62,16 +62,17 @@ Set `SEED_DIR` to this skill directory (path containing `SKILL.md`).
 
 Default corpus is the offline **fixture** (5 rows) under `providers/matraix-persona-1m/fixtures/`. `capabilities()` / provenance report `corpusMode: "fixture"` and a fixture dataset id — not the Hugging Face 1M release.
 
-To use a larger decoded corpus:
+Larger corpora (per provider):
 
 ```bash
-# optional download + decode first N rows (needs python3 + pyarrow [+ huggingface-cli])
+# MatrAIx — optional download + decode (python3 + pyarrow [+ huggingface-cli])
 bash ${SEED_DIR}/scripts/decode-matraix.sh --download --limit 1000
-export MATRAIX_CORPUS_PATH=./matraix-1m.decoded.json   # or .jsonl
-node ${SEED_DIR}/scripts/prepare-corpus.js --validate "$MATRAIX_CORPUS_PATH" --limit 20
-```
+export MATRAIX_CORPUS_PATH=./matraix-1m.decoded.json
 
-Decoder: `providers/matraix-persona-1m/scripts/decode_parquet.py`
+# Nemotron Personas USA — export HF rows to JSON/JSONL, then:
+export NEMOTRON_CORPUS_PATH=./nemotron-usa.jsonl
+node ${SEED_DIR}/scripts/search.js --repo nvidia/Nemotron-Personas-USA --intent '{"domain":["software"]}'
+```
 
 Empty filter results return `[]` (no silent unfiltered fallback). Re-check intent if search is empty.
 

diff --git a/skills/persona-seed/providers/README.md b/skills/persona-seed/providers/README.md
@@ -7,6 +7,9 @@ Adapters that turn corpus rows into `SeedProfile`. Orchestration: `../SKILL.md`.
 | Provider id | Family | Status | HF repos |
 |-------------|--------|--------|----------|
 | `matraix-persona-1m` | attribute-census | ga | `MatrAIx2026/MatrAIx_Persona_1M` |
+| `nemotron-personas-usa` | attribute-census | ga | `nvidia/Nemotron-Personas-USA` |
+
+Keep `seed-capable.public.json` in sync when adding `hfRepos` (mirror into the frontend repo’s `lib/persona-seed-capable.json` for `/datasets` badges).
 
 `openpersona.co/datasets` may list many more corpora. Only registry rows are usable with persona-seed.
 

diff --git a/skills/persona-seed/providers/nemotron-personas-usa/fixtures/sample-corpus.json b/skills/persona-seed/providers/nemotron-personas-usa/fixtures/sample-corpus.json
@@ -0,0 +1,127 @@
+[
+  {
+    "uuid": "nemo-fixture-001",
+    "persona": "A pragmatic mid-career software engineer who prefers clear specs, short feedback loops, and mentorship over heroics.",
+    "professional_persona": "Backend engineer focused on reliability and API design.",
+    "sports_persona": "Weekend trail runner.",
+    "arts_persona": "Amateur photographer.",
+    "travel_persona": "Plans trips around national parks.",
+    "culinary_persona": "Home cook who meal-preps on Sundays.",
+    "cultural_background": "First-generation American; bilingual household growing up.",
+    "skills_and_expertise": "Distributed systems, PostgreSQL, incident response.",
+    "skills_and_expertise_list": "python;sql;kubernetes;incident response",
+    "hobbies_and_interests": "Trail running, photography, science fiction.",
+    "hobbies_and_interests_list": "trail running;photography;sci-fi",
+    "career_goals_and_ambitions": "Grow into a staff engineer role while mentoring juniors.",
+    "sex": "Female",
+    "age": 34,
+    "marital_status": "Married",
+    "education_level": "Master's",
+    "bachelors_field": "Computer Science",
+    "occupation": "Software engineer",
+    "city": "Seattle",
+    "state": "WA",
+    "zipcode": "98101",
+    "country": "USA"
+  },
+  {
+    "uuid": "nemo-fixture-002",
+    "persona": "A patient high-school history teacher who asks Socratic questions and keeps language plain.",
+    "professional_persona": "Public school educator specializing in US history.",
+    "sports_persona": "Coach of the school debate team more than athletics.",
+    "arts_persona": "Community theater volunteer.",
+    "travel_persona": "Museum-first city trips.",
+    "culinary_persona": "Potluck organizer.",
+    "cultural_background": "Midwest upbringing; values civic engagement.",
+    "skills_and_expertise": "Curriculum design, classroom facilitation.",
+    "skills_and_expertise_list": "curriculum;facilitation;public speaking",
+    "hobbies_and_interests": "History podcasts, coaching, gardening.",
+    "hobbies_and_interests_list": "podcasts;coaching;gardening",
+    "career_goals_and_ambitions": "Become a department lead and write a teaching handbook.",
+    "sex": "Male",
+    "age": 41,
+    "marital_status": "Married",
+    "education_level": "Bachelor's",
+    "bachelors_field": "History",
+    "occupation": "High school teacher",
+    "city": "Columbus",
+    "state": "OH",
+    "zipcode": "43215",
+    "country": "USA"
+  },
+  {
+    "uuid": "nemo-fixture-003",
+    "persona": "A cautious clinical research coordinator who hedges claims and refuses medical authority cosplay.",
+    "professional_persona": "Clinical research coordinator at an academic medical center.",
+    "sports_persona": "Yoga twice a week.",
+    "arts_persona": "Classical piano.",
+    "travel_persona": "Conference travel only when necessary.",
+    "culinary_persona": "Vegetarian home cooking.",
+    "cultural_background": "Immigrant family; strong emphasis on education.",
+    "skills_and_expertise": "Protocol compliance, patient scheduling, literature review.",
+    "skills_and_expertise_list": "GCP;literature review;scheduling",
+    "hobbies_and_interests": "Biostatistics blogs, classical music.",
+    "hobbies_and_interests_list": "biostats;classical music",
+    "career_goals_and_ambitions": "Move into clinical operations management.",
+    "sex": "Female",
+    "age": 48,
+    "marital_status": "Single",
+    "education_level": "PhD",
+    "bachelors_field": "Biology",
+    "occupation": "Clinical research coordinator",
+    "city": "Boston",
+    "state": "MA",
+    "zipcode": "02115",
+    "country": "USA"
+  },
+  {
+    "uuid": "nemo-fixture-004",
+    "persona": "An energetic indie e-commerce seller who ships fast experiments and talks in short upbeat paragraphs.",
+    "professional_persona": "Shopify store operator and brand builder.",
+    "sports_persona": "Pick-up basketball.",
+    "arts_persona": "Product photography and Canva design.",
+    "travel_persona": "Wholesale market trips.",
+    "culinary_persona": "Street-food explorer.",
+    "cultural_background": "West Coast creative community.",
+    "skills_and_expertise": "Social commerce, creative ads, customer support.",
+    "skills_and_expertise_list": "shopify;canva;meta ads",
+    "hobbies_and_interests": "Design, street food, community markets.",
+    "hobbies_and_interests_list": "design;street food;markets",
+    "career_goals_and_ambitions": "Grow a lifestyle brand to seven figures.",
+    "sex": "Male",
+    "age": 23,
+    "marital_status": "Single",
+    "education_level": "Some college",
+    "bachelors_field": "",
+    "occupation": "E-commerce seller",
+    "city": "Los Angeles",
+    "state": "CA",
+    "zipcode": "90012",
+    "country": "USA"
+  },
+  {
+    "uuid": "nemo-fixture-005",
+    "persona": "A structured product manager who frames tradeoffs clearly and stays oriented to user outcomes.",
+    "professional_persona": "B2B SaaS product manager.",
+    "sports_persona": "Indoor climbing.",
+    "arts_persona": "Design critique hobbyist.",
+    "travel_persona": "Customer onsite visits.",
+    "culinary_persona": "Coffee snob.",
+    "cultural_background": "Suburban Northeast; pragmatic communicator.",
+    "skills_and_expertise": "Roadmapping, UX research synthesis, SQL for analytics.",
+    "skills_and_expertise_list": "roadmap;ux research;sql;notion",
+    "hobbies_and_interests": "Systems thinking, climbing, espresso.",
+    "hobbies_and_interests_list": "systems thinking;climbing;coffee",
+    "career_goals_and_ambitions": "Lead a product org focused on developer tools.",
+    "sex": "Female",
+    "age": 29,
+    "marital_status": "Partnered",
+    "education_level": "Bachelor's",
+    "bachelors_field": "Economics",
+    "occupation": "Product manager",
+    "city": "New York",
+    "state": "NY",
+    "zipcode": "10001",
+    "country": "USA"
+  }
+]
diff --git a/skills/persona-seed/providers/nemotron-personas-usa/mapping.md b/skills/persona-seed/providers/nemotron-personas-usa/mapping.md
@@ -0,0 +1,27 @@
+# nemotron-personas-usa → SeedProfile
+
+Source: [`nvidia/Nemotron-Personas-USA`](https://huggingface.co/datasets/nvidia/Nemotron-Personas-USA)  
+Family: `attribute-census`
+
+## Field map
+
+| SeedProfile | Nemotron field(s) |
+|-------------|-------------------|
+| `identity.summary` | `persona` (fallback: `professional_persona`) |
+| `identity.occupation` | `occupation` |
+| `identity.education` | `education_level` (+ `bachelors_field` when present) |
+| `identity.region` | `city`, `state`, `country` joined |
+| `identity.ageBracket` | derived from `age` |
+| `identity.locale` | `["en"]` for USA rows |
+| `character.traits` | light keywords from persona texts (optional) |
+| `character.interests` | `hobbies_and_interests_list` (split on `;` / `,`) |
+| `character.motivations` | `career_goals_and_ambitions` (single entry) |
+| `capabilities.skills` / `tools` | `skills_and_expertise_list` |
+| `capabilities.domains` | occupation / bachelor field tokens |
+| `evidence.rawDescriptions` | persona + professional_persona + cultural_background |
+| `constraints.sensitiveFlags` | `healthcare_domain` when occupation/persona mentions clinical/medical |
+
+## Corpus
+
+- Default: offline fixture (5 rows)
+- External: `NEMOTRON_CORPUS_PATH` → JSON array or `.jsonl` of raw Nemotron rows