|
|
| https://huggingface.co/spaces/decodingdatascience/newrag-pine |
|
|
| # 📞 Telecom Customer Support LLM with Groq API |
|
|
| This project demonstrates how to build a fast, production-grade AI-powered telecom customer support assistant using the **Groq API** and optimized GenAI configurations. |
|
|
| ## 📌 Project Overview |
|
|
| A step-by-step guide to: |
| - Sending POST requests using **Postman** |
| - Connecting to the **Groq API** |
| - Testing default vs. optimized GenAI configurations |
| - Applying structured **prompt templates** |
| - Deploying a simple LLM-powered support API |
|
|
| ## 🔧 Tech Stack |
|
|
| | Layer | Tool/Tech | |
| |---------------|-------------------------| |
| | LLM | [Groq API](https://groq.com/) | |
| | API Platform | FastAPI / Postman | |
| | Prompt Design | Custom templates | |
| | Deployment | Localhost / Cloud (optional) | |
|
|
| ## 🧠 AI Configuration |
|
|
| | Parameter | Description | |
| |---------------------|--------------------------------------| |
| | `temperature` | Controls randomness (default: 0.7) | |
| | `top_p` | Nucleus sampling | |
| | `max_tokens` | Max tokens to generate | |
| | `frequency_penalty` | Repetition control | |
| | `presence_penalty` | Topic diversity | |
|
|
| ## 🔄 Experiment Setup |
|
|
| ### 1. No Prompt Template + Default Config |
| - Basic user input |
| - Uses Groq defaults |
| - For benchmarking |
|
|
| ### 2. With Prompt Template + Tuned Config |
| - Structured input (e.g., role, intent, constraints) |
| - Custom temperature and token limits |
| - Optimized for domain-specific responses |
|
|
| ## 🚀 Quickstart |
|
|
| ### Step 1: Clone the repo |
|
|
| ```bash |
| git clone https://github.com/your-username/telecom-support-llm.git |
| cd telecom-support-llm |
| |
