Muhammad Fayaz Khan
LLM Engineer | Transforming Your Ideas Into AI Products
New to Contra
Muhammad Fayaz is building their profile!
UrduGPT is an end-to-end LLM engineering project that demonstrates a complete post-training workflow, featuring supervised fine-tuning (SFT) and Direct Preference Optimization (DPO) for the Urdu language. Utilizing the Hugging Face ecosystem, PEFT, and Unsloth, the project delivers highly aligned models for better linguistic performance and enables LLM-as-a-Judge evaluation via Groq. The complete project, including SFT and DPO models, is available for review on GitHub and Hugging Face.
0
0
AI-powered business document assistant built with RAG pipeline.
What it does
Upload your business PDFs and ask questions in plain language. Get accurate answers with source citations instantly.
Tech Stack:
FastAPI —> backend
Chainlit —> chat UI
ChromaDB —> vector database
Groq (Llama 3) —> LLM
HuggingFace —> embeddings
Cohere Rerank —> reranking
BM25 —> hybrid search
0
0
CompetiTrack AI is an AI-powered competitor monitoring system that tracks product price and stock changes on e-commerce websites. It detects meaningful updates, generates business insights using an LLM, and sends email notifications only when changes occur
Keeping track of competitor pricing manually is time-consuming and difficult to scale. CompetiTrack AI automates the entire process.
The system scrapes competitor websites, compares the latest data with the previous snapshot, analyzes the changes using an LLM, and sends an email report only when something has changed. This avoids unnecessary notifications while providing useful insights for business decisions.
Features
- Competitor website monitoring
- Product price tracking
- Stock availability tracking
- Change detection
- AI-generated business insights
- Email notifications
- FastAPI backend
- Streamlit dashboard
- LangGraph workflow
- Playwright-based web scraping
2
0
4