Spaces:
Running
Running
| title: README | |
| emoji: ⚡ | |
| colorFrom: red | |
| colorTo: indigo | |
| sdk: static | |
| pinned: false | |
| # Codegnan IT Solutions | |
| **Building and running self-hosted AI for Indian languages ** | |
| 🌐 [Website](https://codegnan.com/) · 💼 [LinkedIn](linkedin.com/company/codegnan?originalSubdomain=in) · ✉️ [[CONTACT_EMAIL]](mailto:[hai@codegnan.com]) | |
| --- | |
| ## About us | |
| Codegnan IT Solutions builds, deploys, and operates AI systems on our own infrastructure, | |
| using open-weight models. Our focus is practical AI for Indian users: Telugu and other Indic | |
| languages, real-world audio, and production reliability over demos. | |
| We also run AI training programs for students, professionals, and business teams. | |
| ## What we work on | |
| | Area | Focus | | |
| |------|-------| | |
| | 🎙️ **Indic speech** | Speech-to-text for Telugu and other Indian languages, including noisy phone-call audio | | |
| | 🗣️ **Voice agents** | Conversational AI agents that speak naturally in Indian languages | | |
| | 📄 **Document intelligence** | OCR and document reconstruction for English and Indic scripts | | |
| | 🧠 **LLM operations** | Serving, monitoring, and governing self-hosted open-weight LLMs | | |
| | 💼 **AI hiring** | — resume matching, ATS analysis, quizzes, and AI interviews | | |
| | 🎓 **Training** | Claude and AI bootcamps for business professionals and developers | | |
| ## How we work | |
| - **Self-hosted first** — models run on our own GPUs; data stays in-house. | |
| - **Open-weight models** — we credit and respect the licenses of the models we build on. | |
| - **Indic-first** — every system is tested on Indian languages, not just English. | |
| - **Measured, not assumed** — we publish numbers only when we have benchmarked them. | |
| ## What you'll find here | |
| | Type | Status | | |
| |------|--------| | |
| | 🤖 **Models** | Coming soon | | |
| | 📊 **Datasets** | Coming soon | | |
| | 🚀 **Spaces** | Coming soon | | |
| ## For members | |
| - Create repos **private by default**; make them public only after review. | |
| - Never upload customer data, call recordings, API keys, or `.env` files. | |
| - Every public model or dataset needs a card with its license and the base model it builds on. | |
| - Naming: `[area]-[project]-[version]` — e.g. `speech-telugu-stt-v1`. | |
| --- | |
| <sub>© Codegnan IT Solutions</sub> |