Author: John Mitchell (@whmatrix) Status: ENTRY_ROUTER (Portfolio Hub)

Professional portfolio showcasing production-grade semantic indexing and RAG infrastructure.

Start Here

Resource	Description
Universal Protocol v4.23	Canonical spec: deliverable contracts, audit contracts, quality gates
semantic-indexing-batch-02	Production implementation: 8.3M+ vectors indexed
research-corpus-discovery	Applied methodology: semantic search across 10 research institutions, 4,600+ documents

Note: semantic-indexing-batch-01 is a superseded foundation batch.

Structural Analysis

• comparative-grammar-gpt-vs-claude — GPT vs Claude dialogue grammar comparison
• structural-collaboration-primitives — Interaction primitive definitions
• interaction-mechanics-index — Dual FAISS indices for dialogue mechanics retrieval

---

whmatrix.github.io

Live site: https://whmatrix.github.io

Production Stats

• Total Vectors: 9,016,688 across 9 production datasets (batch-01 + batch-02)
• Research Corpora: 4,600+ documents across 10 institutions (~75,000 chunks)
• Datasets: Wikipedia Featured Articles, StackExchange Python, ArXiv ML Abstracts, 20 Newsgroups, SimpleWiki, IMDB, StackOverflow, AG News, Disaster Tweets
• Model: intfloat/e5-large-v2 (1024-dim embeddings)
• Index: FAISS IndexFlatIP

---

Visual Overview

+-------------------------------------------------------+
|            whmatrix: Semantic Search Stack             |
+-------------------------------------------------------+
|                                                       |
|  SCALE          STANDARDS     THEORY     APPLICATION  |
|    |               |            |            |        |
|    +- 661K       +- Protocol  +- Grammar  +- 10      |
|    +- 8.35M      +- v4.23    +- Primitives  orgs    |
|                               +- Mechanics           |
|                                                       |
|  [Try Demo]  [View Specs]  [See Architecture]         |
+-------------------------------------------------------+

Vectors indexed:      9,016,688 (across 9 datasets)
Documents analyzed:   4,600+ (across 10 institutions)
Query latency:        <100ms (single query)
Embedding model:      e5-large-v2 (1024-dim, FP16)
Index type:           FAISS IndexFlatIP

---

Scope

This site serves as a presentation and routing layer, not a full technical explanation.

• Static HTML/CSS only
• No JavaScript frameworks
• No blog or research log
• No client-specific material

All production guarantees, acceptance criteria, and verification artifacts are defined in the protocol repository.