Create README.md
Browse files
README.md
CHANGED
|
@@ -1,10 +1,113 @@
|
|
| 1 |
---
|
| 2 |
-
|
| 3 |
-
|
| 4 |
-
colorFrom: purple
|
| 5 |
-
colorTo: green
|
| 6 |
sdk: static
|
| 7 |
-
|
|
|
|
| 8 |
---
|
| 9 |
|
| 10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
title: Axiveri
|
|
|
|
|
|
|
| 4 |
sdk: static
|
| 5 |
+
emoji: π
|
| 6 |
+
short_description: 'African AI Infrastructure β Building intelligent systems '
|
| 7 |
---
|
| 8 |
|
| 9 |
+
<div align="center">
|
| 10 |
+
|
| 11 |
+
# AXIVERI
|
| 12 |
+
### African AI Infrastructure
|
| 13 |
+
|
| 14 |
+
*We build AI systems designed for the African context β*
|
| 15 |
+
*its infrastructure, its languages, its developer realities,*
|
| 16 |
+
*and its regulatory environment.*
|
| 17 |
+
|
| 18 |
+
---
|
| 19 |
+
|
| 20 |
+
</div>
|
| 21 |
+
|
| 22 |
+
## The Problem
|
| 23 |
+
|
| 24 |
+
Most AI models are trained on internet data dominated by Western contexts.
|
| 25 |
+
|
| 26 |
+
They know what a dollar is. They do not know what kobo is.
|
| 27 |
+
They can write a Stripe integration. They cannot write a NIBSS NIP integration.
|
| 28 |
+
They understand GDPR. They have never heard of NDPR.
|
| 29 |
+
|
| 30 |
+
700,000 software developers in Nigeria alone.
|
| 31 |
+
A fintech sector processing billions of transactions annually.
|
| 32 |
+
Zero AI systems built specifically for them.
|
| 33 |
+
|
| 34 |
+
This is not a gap we intend to complain about.
|
| 35 |
+
**It is a gap we are building to close.**
|
| 36 |
+
|
| 37 |
+
---
|
| 38 |
+
|
| 39 |
+
## What We Build
|
| 40 |
+
|
| 41 |
+
Axiveri develops domain-specific AI model series for Africa β starting with language models and expanding into health, security, legal, and financial intelligence.
|
| 42 |
+
|
| 43 |
+
Our first model series is **Africlaude**.
|
| 44 |
+
|
| 45 |
+
Africlaude models are fine-tuned large language models trained on curated knowledge covering:
|
| 46 |
+
|
| 47 |
+
- Nigerian fintech engineering (Paystack, Flutterwave, NIBSS, kobo/naira)
|
| 48 |
+
- CBN regulatory compliance and KYC frameworks
|
| 49 |
+
- AML, fraud detection, and transaction monitoring
|
| 50 |
+
- USSD architecture and Africa's Talking integration
|
| 51 |
+
- African language contexts β Pidgin English, Yoruba, Hausa, Igbo
|
| 52 |
+
- Security engineering for African infrastructure
|
| 53 |
+
|
| 54 |
+
---
|
| 55 |
+
|
| 56 |
+
## Africlaude β Coming Soon
|
| 57 |
+
|
| 58 |
+
**Africlaude** is our flagship model series β a family of language models fine-tuned specifically for African developers, engineers, and professionals.
|
| 59 |
+
|
| 60 |
+
**Africlaude 7B** is the first release.
|
| 61 |
+
|
| 62 |
+
> Trained across 9 cycles on Tesla T4 GPU infrastructure.
|
| 63 |
+
> Evaluated on AfriclaudeBench v1.0 β a benchmark we designed
|
| 64 |
+
> specifically to measure AI competency in African fintech and developer domains.
|
| 65 |
+
|
| 66 |
+
π¬ **Follow this organisation to be notified on release.**
|
| 67 |
+
|
| 68 |
+
---
|
| 69 |
+
|
| 70 |
+
## Our Approach
|
| 71 |
+
|
| 72 |
+
We believe the next generation of AI for Africa will not come from importing models built elsewhere and hoping they work here.
|
| 73 |
+
|
| 74 |
+
It will come from organisations embedded in the African context β who understand what it means to build for 2G networks, USSD sessions, agent banking, CBN regulatory frameworks, and a developer community growing faster than any on earth.
|
| 75 |
+
|
| 76 |
+
We are building that infrastructure.
|
| 77 |
+
**Open. African. Rigorous.**
|
| 78 |
+
|
| 79 |
+
---
|
| 80 |
+
|
| 81 |
+
## Organisation Structure
|
| 82 |
+
Axiveri
|
| 83 |
+
βββ Africlaude (Model Series)
|
| 84 |
+
βββ Africlaude 7B β Language model for fintech & coding
|
| 85 |
+
βββ Africlaude 14B β In development
|
| 86 |
+
βββ Africlaude-* β Domain-specific models (coming)
|
| 87 |
+
---
|
| 88 |
+
|
| 89 |
+
## Partnerships & Collaboration
|
| 90 |
+
|
| 91 |
+
We are open to research collaborations, dataset partnerships, and integration agreements with:
|
| 92 |
+
|
| 93 |
+
- African financial institutions and fintechs
|
| 94 |
+
- Healthcare providers and hospital networks
|
| 95 |
+
- Technology companies and developer platforms
|
| 96 |
+
- Academic institutions and research organisations
|
| 97 |
+
- Investors aligned with African technology infrastructure
|
| 98 |
+
|
| 99 |
+
---
|
| 100 |
+
|
| 101 |
+
<div align="center">
|
| 102 |
+
|
| 103 |
+
**Founder:** Emmanuel Ariyo β Axiveri, Nigeria
|
| 104 |
+
|
| 105 |
+
[
|
| 106 |
+
|
| 107 |
+

|
| 108 |
+
|
| 109 |
+
](https://x.com/AfriclaudeAI)
|
| 110 |
+
|
| 111 |
+
*Building the AI infrastructure Africa deserves.*
|
| 112 |
+
|
| 113 |
+
</div>
|