Add AI Engineer archetype

[SaharPak]

Why

The README advertises 6 archetypes (AI Product Manager, AI Engineer, Backend, Frontend, DevOps, Data Engineer) but only 3 are implemented. Of the missing ones, AI Engineer is the most-requested role in the 2026 market.

The current data-ml-engineer archetype is tuned to the legacy data stack (Python, SQL, Spark, Airflow, dbt, pandas, PyTorch, TensorFlow) — it doesn't capture the LLM/agent stack that defines AI Engineering today (RAG, MCP, Claude, Cursor, embeddings, vector DBs, prompt engineering, agentic patterns, model serving, inference).

What to do

1. Use the New Archetype issue template if it exists, or follow the structure in packages/intelligence/src/archetypes/data-ml-engineer.ts.
2. Research 15–30 keywords from real AI Engineer job postings (LinkedIn, Indeed, Greenhouse public postings). Suggested starting set:
   • RAG, embeddings, vector database (Pinecone, Weaviate, Chroma, Qdrant, pgvector)
   • LLM (Claude, GPT, Llama, Mistral)
   • Prompt engineering, prompt design
   • Agentic patterns, multi-agent systems
   • MCP (Model Context Protocol)
   • LangChain, LlamaIndex, Hugging Face
   • Cursor, Claude Code, dev tools
   • Inference, model serving, vLLM, Triton
   • Fine-tuning, RLHF, evals
   • Python, TypeScript (the two real languages for AI engineers)
3. Define evaluationWeights — start by copying from data-ml-engineer and adjusting.
4. Add to packages/intelligence/src/archetypes/index.ts and packages/intelligence/src/index.ts.
5. Add 5–10 tests to packages/intelligence/src/__tests__/.
6. Submit a PR.

Files to reference

• packages/intelligence/src/archetypes/data-ml-engineer.ts — closest existing example
• packages/intelligence/src/rubric.ts — the 6 rubric dimensions
• packages/schemas/src/archetype.ts — the Zod schema for an archetype

Effort

M (2–4 hours).

Impact

High. Closes the most-requested archetype gap. Lets candidates targeting AI Engineer roles get a real evaluation instead of being default-fallback'd to software-engineer.

[qafariamirhossein]

Hi @SaharPak , I’d like to take this one.

I’ll follow the existing data-ml-engineer archetype structure and create a dedicated AI Engineer archetype focused on the modern LLM/agent stack: RAG, embeddings, vector databases, prompt engineering, MCP, LangChain/LlamaIndex, model serving, inference, evals, Python, and TypeScript.

I’ll also update the required exports in archetypes/index.ts and index.ts, define suitable evaluationWeights, and add tests under packages/intelligence/src/__tests__/ to make sure the new archetype is covered properly.

Please assign this to me if it’s available.