Update README.md

README.md

@@ -1,10 +1,87 @@
 ---
-title: README
-emoji: 🌖
-colorFrom: red
-colorTo: yellow
+title: Typica.ai
+emoji: 🌍
+colorFrom: blue
+colorTo: indigo
 sdk: static
-pinned: false
+pinned: true
 ---
 
-Edit this `README.md` markdown file to author your organization card.
+# Typica.ai
+
+**Research-first NLP for underrepresented languages and specialized domains.**
+
+Typica.ai builds culturally aware, context-specific AI systems — from tokenizers to instruction-tuned language models — with a strong focus on **low-resource languages**, **high-impact domains**, and **linguistically complex environments**.
+
+We believe that AI should not just be **accessible** to underserved communities — it should be **built for and by them**.
+
+---
+
+## 🚀 Our Mission
+
+AI is becoming the infrastructure of the digital age — yet most of it is trained on English-speaking, high-resource contexts.
+
+**Typica.ai exists to correct this imbalance.**
+
+We design, train, and deploy NLP systems that reflect **how people actually speak, think, and live** — in multiple languages, across multiple registers, and rooted in local realities.
+
+We are not fine-tuning global models for surface localization.  
+We are building **original, indigenous AI** that solves real-world problems.
+
+---
+
+## 🛠️ What We Offer
+
+Typica.ai provides expert services in research, model development, and strategic NLP consulting:
+
+- ✅ **AI & NLP Consulting** – Strategy, implementation, and model optimization  
+- ✅ **LLM Fine-Tuning & Alignment** – SFT, DPO, ORPO for low-resource contexts  
+- ✅ **Training & Coaching** – Hands-on education for professionals and teams  
+- ✅ **Data Creation & Curation** – High-quality, task-specific NLP datasets  
+- ✅ **Custom AI Solutions** – Tailored models for specialized use cases
+
+We’re bridging the AI gap for **non-mainstream languages and industries**, delivering smarter, more context-aware AI solutions.
+
+---
+
+## 🧪 Our Work
+
+Our portfolio includes foundational models for:
+
+- Named Entity Recognition (NER)  
+- Sentiment Analysis  
+- Part-of-Speech Tagging  
+- Toxicity Detection  
+- Semantic Search  
+- Multilingual Instruction-Tuned LLMs
+
+All built through reproducible pipelines, evaluated beyond standard benchmarks, and optimized for real-world semantic fidelity.
+
+---
+
+## 🌍 Languages We Work With
+
+We operate across multilingual, code-switched environments — not just between languages, but **within them**, as they are spoken and used:
+
+- 🇲🇦 **Darija** – informal, expressive, identity-rich  
+- 🕌 **Modern Standard Arabic** – formal and institutional  
+- 🇫🇷 **French** – dominant in law, education, and administration  
+- 🌐 **English** – critical for research and global integration
+
+---
+
+## 🤝 Let’s Collaborate
+
+Whether you’re a researcher, developer, educator, or policymaker, we’re ready to collaborate.
+
+📬 Contact us at **contact@typica.ai**  
+🌐 Visit: [https://typica.ai](https://typica.ai)  
+📺 YouTube: [@typica-ai](https://www.youtube.com/@typica-ai2433)  
+📚 Blog: [https://typica.ai/blog](https://typica.ai/blog)
+
+---
+
+**Typica.ai**  
+*Founded by Dr. Hicham Assoudi*  
+Bootstrapped. Research-led. Culture-centered.  
+**AI that belongs. Built with care.**