Skip to content

Proven Results

Case Studies

Real results from sovereign AI infrastructure. Every system described here is live in production, built and operated without third-party gatekeepers.

AI/ML

Sovereign LLM Training

Built and trained the Kush model line, led in production by hotep-llm-kush-v91 (Llama 3.1 8B rsLoRA merged to Q4_K_M GGUF). Fine-tuned across 17+ iterations and 6,266 curated Afrocentric examples (5,271 core + 995 precision entity-recovery SFT), passing 7/7 (100%) on Prime Agent E2E, 7/7 (100%) on the Exact Entity Recall Gate, and 6/6 on safety & coding boundaries at 43.9–48.2 tok/s on local Apple Silicon.

6,266
Training Examples
7/7
Prime & Entity Gate
48.2
Local Metal tok/s
Infrastructure

24/7 Telegram Bot

Deployed @hotep_llm_bot serving real users around the clock with auto-recovery watchdog, sub-2 second response times, persistent user profiles with streak tracking, gamification (XP, badges, streaks), and /deep research mode.

24/7
Uptime
<2s
Response Time
17
Bot Commands
Content

Knowledge Base Platform

Built knowledge.askhotep.ai with 418 static pages across 7 pillars and 437+ indexed RAG documents across 4 vector collections — protocols, history, wisdom, and entities. Static-generated for instant page loads with semantic search via RAG.

437+
RAG Documents
4
Collections
418
Static Pages
Business

Revenue Pipeline

End-to-end monetization with Stripe integration, webhook gateway on Cloudflare Tunnel, 2 pricing tiers (Free + Pro $9/mo), Telegram Stars payment support, and cache-first checkout.

2
Pricing Tiers
Live
Webhooks
0ms
Cache-First Checkout
DevOps

Zero-Cost Infrastructure

Production inference runs on owned local hardware with Cloudflare Pages for static hosting and Cloudflare Tunnel for API proxy. Total monthly cloud cost: $0.

$0/mo
Cloud Cost
Mac Mini
Self-Hosted Runtime
99.9%
Uptime Target
AI/ML

Advanced RAG Pipeline

Query expansion generates 3 search variants per query. LLM + cross-encoder dual reranking for precision. Parallel search across 4 ChromaDB collections. Redis embedding cache for 85%+ cache hit rate. 5 Prometheus metrics for real-time monitoring.

437+
Documents
85%+
Cache Hit Rate
5
Prometheus Metrics

Ready to build sovereign AI?

Experience Hotep Intelligence firsthand. Try the live demo or explore our pricing to find the right tier for your needs.