added the readme file

README-ammas.md

@@ -0,0 +1,58 @@
+# Ahmed Sharaf
+
+Hi there! 👋 I'm Ahmed, a passionate software engineer and researcher with a strong background in Artificial Intelligence, NLP, and cutting-edge data visualization. I thrive on building impactful projects and solving challenging problems. Here's a glimpse of my work and skills:
+
+---
+
+## 🌟 Projects
+
+### 🔍 **VSR (Lip Reading) for Egyptian Arabic**
+- **Tech Stack:** Python, TensorFlow, OpenCV, GCP, FFmpeg  
+- I spearheaded the creation of the **first-ever Visual Speech Recognition (VSR) dataset for Egyptian Arabic**, working with a team of 33 volunteers.  
+- Developed 7 state-of-the-art VSR models with remarkable accuracy metrics:
+  - **Top-1 Accuracy:** 97.44%  
+  - **Character Error Rate:** as low as 3.8%  
+- This project broke new ground for VSR in low-resource languages, opening doors for advanced accessibility and communication technologies in the region.
+
+---
+
+## 💼 Professional Experience
+
+### 💡 **Software Engineer Intern at Valuenex Inc.**
+*Jun 2022 – Sep 2024 | Palo Alto, CA*  
+- Developed a **document categorization algorithm** using computational geometry and hierarchical data mapping.  
+  - It became a **core production feature**, adopted by **half of the Fortune 500 companies in Japan**.  
+- Designed and built a **Unity 3D game** that landed the company its **largest-ever contract** with Otsuka, a leading pharmaceutical company ($30Bn+ market cap).  
+- Created a **GPU-accelerated interface** for projecting 5M+ data points as documents on a spherical surface.  
+  - Users could explore immersive 360° visualizations via VR or mobile devices.  
+  - Optimized interaction with **precomputation and boundary-aware caching**, achieving a **250x speedup**.  
+- Launched a **mental health app** for clinical trials:
+  - Built with MongoDB, RESTful APIs, and React Native.  
+  - A/B tested and deployed multiple iterations to maximize user engagement and stress reduction.
+
+---
+
+### 🧠 **Stanford NLP Researcher (Prof. Monica Lam & Chris Manning)**  
+- **GPT Agents and Retrieval-Augmented Generation (RAG):**  
+  - Integrated GPT feedback agents into **STORM**, a multi-agent LLM generating comprehensive, cited articles for **30k+ users**.  
+  - Built and deployed a **custom RAG pipeline** using Qdrant and Langchain, enabling personalized data retrieval from user-uploaded repositories.  
+- **Advancing NLP for Arabic:**  
+  - Fine-tuned Google’s **BERT model** on 2.5M+ Egyptian Arabic sentences, achieving **67.2% F1 score** on unsupervised question answering.  
+  - Developed tools for annotating and training Egyptian Arabic NLP models:
+    - Improved NER accuracy from **74% to 88.6%** with a Flask annotation app.  
+
+---
+
+## 🛠️ Technical Skills
+
+- **Languages:** Python, C, C++, C#, JavaScript, Swift, SQL  
+- **Frameworks/Tools:** React Native, Node.js, Flask, Firebase, MySQL, MongoDB  
+- **Libraries:** PyTorch, TensorFlow, Keras, OpenCV, Scikit-learn, Pandas, Matplotlib  
+
+---
+
+## 🎖️ Achievements
+
+- **Egypt National Valedictorian** (Rank: 1/200,000)  
+- Competitor in the **International Math Olympiad** (2x) and **Intel ISEF** (2x)  
+- Finalist in the **Egypt Physics Olympiad**