Upload README.md with huggingface_hub

README.md

@@ -9,51 +9,30 @@ pinned: false
 
 # MainStack
 
-**Specializing in ServiceNow Agentic Delivery.**
+### ServiceNow Agentic Delivery — and the open models behind it.
 
-MainStack builds AI systems and open models for the full ServiceNow delivery
-lifecycle — from business analysis and solution design through user stories,
-implementation, testing, and validation.
+[mainstack.co.uk](https://www.mainstack.co.uk/) · [LinkedIn](https://linkedin.com/company/mainstack-it)
 
-## 🚀 Featured release
+MainStack is a consultancy specializing in ServiceNow Agentic Delivery. We build the AI that does the drafting work of a delivery engagement — and we release it openly. **marvy-14B** is our first public model: a fine-tuned LLM that drafts artifacts across the entire ServiceNow delivery lifecycle, from business analysis through Solution Design Documents, user stories, test cases, and validation. It's a first-draft specialist built for the people who do the work — solution architects, business analysts, technical consultants, project managers, and delivery leads.
 
-### [marvy-14B](https://huggingface.co/MainStack/marvy-14B) — the first open fine-tuned model for full-cycle ServiceNow delivery
-
-`marvy-14B` is a fine-tune of [Qwen2.5-14B-Instruct](https://huggingface.co/Qwen/Qwen2.5-14B-Instruct)
-(Apache-2.0) specialized for end-to-end ServiceNow delivery work. It was trained
-on a curated, fully anonymized corpus of real engagement artifacts spanning the
-delivery chain practiced by our SN-Architect operators:
-
-> project organizer → business-analyst → solutions-architect → project manager → technical-consultant → validation-architect
-
-It produces implementation-grade first drafts of:
+## 🔒 Built responsibly
 
-| | |
-|---|---|
-| 📋 Business analyses | 📐 Solution Design Documents (SDD) |
-| ✅ Requirements & acceptance criteria | 📝 User stories |
-| 👥 Stakeholder maps | 🔧 Implementation plans (tables, plugins, sys_ids) |
-| 🗂️ Systems inventories | 🧪 Test cases |
-| 🔍 Validation & gap analysis | 🔗 End-to-end delivery chains (story → build → test) |
+marvy was trained on real engagement artifacts, which means privacy was a design constraint, not an afterthought. Every training artifact was rigorously anonymized and redacted to **zero residual PII**, verified by an automated leakage scanner. Evaluation was run on a **project- and customer-disjoint** held-out split, so the reported perplexity reflects genuine generalization to unseen work — not memorization. We frame marvy honestly: it accelerates first drafts; it does not replace the judgment of a consultant, and it is not a tool-use or agentic fine-tune.
 
-**Available formats:**
+## 📦 Available formats
 
-- [`MainStack/marvy-14B`](https://huggingface.co/MainStack/marvy-14B) — merged FP16 (Transformers / vLLM / MLX)
-- [`MainStack/marvy-14B-lora`](https://huggingface.co/MainStack/marvy-14B-lora) — LoRA adapter (~175 MB)
-- [`MainStack/marvy-14B-GGUF`](https://huggingface.co/MainStack/marvy-14B-GGUF) — Q4_K_M & Q8_0 quants for Ollama / llama.cpp / LM Studio
+| Repo | Format | Use case |
+|------|--------|----------|
+| [MainStack/marvy-14B](https://huggingface.co/MainStack/marvy-14B) | Merged FP16 | Full-precision inference and serving |
+| [MainStack/marvy-14B-lora](https://huggingface.co/MainStack/marvy-14B-lora) | LoRA adapter | Compose on top of Qwen2.5-14B-Instruct |
+| [MainStack/marvy-14B-GGUF](https://huggingface.co/MainStack/marvy-14B-GGUF) | GGUF quants | Local runs on Apple Silicon, LM Studio, Ollama |
 
 ```bash
 ollama run hf.co/MainStack/marvy-14B-GGUF:Q4_K_M
 ```
 
-## 🔒 Built responsibly
-
-Our training corpus is rigorously anonymized: customer/partner names aliased,
-emails/hostnames/IPs scrubbed, credential files excluded, and a leakage scanner
-that asserts **zero** residual PII in training content.
-
-## 🌐 About
+## About
 
-MainStack — [mainstack.co.uk](https://www.mainstack.co.uk/) · [LinkedIn](https://linkedin.com/company/mainstack-it)
+MainStack specializes in ServiceNow Agentic Delivery, building open, fine-tuned models that draft the artifacts of a delivery engagement — and releasing them to the practitioners who use them. Learn more at [mainstack.co.uk](https://www.mainstack.co.uk/).
 
 *Released under Apache-2.0. Built with Qwen.*