Mohamed Ahmed

Developed an AI-powered personalized learning platform using LLAMA3, Retrieval-Augmented Generation (RAG), Knowledge Graphs, and Streamlit. The platform generates adaptive study roadmaps, personalized educational content, and interactive quizzes based on real-time learner assessments

Built an automated n8n workflow to process invoice PDFs from Google Drive using a local LLM via LM Studio. Extracted structured invoice data including invoice number, client details, total amount, invoice date, and due date. Stored extracted data in Google Sheets and automated email notifications through Gmail.

Built a Retrieval-Augmented Generation (RAG) workflow using self-hosted n8n that ingests documents from Google Drive, extracts and chunks text with a Recursive Character Text Splitter, generates embeddings locally using LM Studio (OpenAI- compatible API), stores vectors in ChromaDB, and enables document-grounded question answering through semantic vector search.

Built an end-to-end AI automation workflow using self-hosted n8n that reads topics from Google Sheets, performs live web research through the Tavily API, and generates structured LinkedIn content using LLMs via OpenRouter. The workflow automatically formats and writes the generated content back to Google Sheets, significantly reducing manual research and content drafting

Developed an AI-powered Gmail automation using self-hosted n8n that classifies incoming emails with LLMs, applies labels automatically, and generates context- aware replies with conversational memory to streamline email management

Built an NLP sentiment analysis model using Support Vector Machine (SVM), achieving 87% classification accuracy through feature engineering and hyperparameter tuning.