Abdulrhman S.

I help businesses solve real-world problems with AI—specializing in Natural Language Processing (NLP) and Machine Learning. From chatbots to document automation, I’ve built 50+ solutions that cut costs, boost efficiency, and scale operations. Explore my portfolio to see the impact. I specialize more in these areas: Semantic Search, Vector Spaces, n8n workflows, Automation Technologies, Langchain Indexing, RAG Systems, Agentic Flows, AI Agents, Vector Databases, LlamaIndex, Ollama, Docling, Tika, OCR, Open Source, Open WebUI, MLOps, FAISS, Azure Data Factory, AWS Glue, ReAct Agents, RASA, chatbot development, Knowledge Store, Knowledge AI ## Projects Portfolio ## ## LLMs & AI Projects ### Text Generation & Language Models - ** E.ON validation RAG System ** - Fine-tuned and Prompt Engineered GPT-4 for extraction of contract prices from company PDF documents - Implemeneted a Langchain based RAG System for automated tabu- lation - Developed an advanced querying system on Databricks to perform Information Retrieval Vector search - Developed an integrated system for processing the documents on Databricks - Created an automated Azure blob storage for the information extracted on Azure - **Lyrics Generation System** - Fine-tuned GPT-3.5 Turbo for creative text generation - Implemented custom rhyme scheme constraints - Developed syllabic pattern enforcement - **Automated Code Generation Platform** - Leveraged LLM fine-tuning for code transformation - Implemented rule-based code generation systems - **Custom Content Generation Applications** - Developed exam question generation system using HuggingFace models - Created personalized email generation system using GPT-3 and LinkedIn data - Implemented marketing content automation ### Language Understanding & Analysis - **Customer Feedback Analysis System** - Built personality analysis components - Implemented sentiment analysis for written text - Developed comprehensive feedback processing pipeline - **Text Simplification Engine** - Utilized tokenized BART model - Integrated Complex Word Identification (CWI) - Focused on accessibility for users with disabilities - **Knowledge Graph QA System** - Implemented adversarial attack resistance - Developed graph-based query processing - Created robust answer generation mechanism ### Speech & Audio Processing - **German Voice Recognition Web Application** - Built with Flask, BERT, JavaScript, HTML, CSS - Implemented real-time speech processing - Created intuitive user interface - **Multilingual Speech Recognition** - Fine-tuned wav2vec transformers for Turkish, Hebrew, and Romanian - Optimized for multiple language support - Implemented acoustic model adaptations - **Arabic Poem Meter Classification** - Developed CNN-based acoustic analysis - Created specialized poetry processing pipeline ## Web Applications & User Interfaces - **Enhanced Chatbot System** - Implemented question type classification - Developed advanced keyword detection - Created comprehensive QA pipeline - **Face Positioning System** - Developed ID card processing application - Built with Flask, HTML, CSS, and Djinga - Implemented precise facial alignment algorithms - **Stock Photography Platform** - Integrated Neural Image Captioning - Utilized CNN-RNN architecture - Processed Google Flickr 8k dataset ## Machine Learning & Computer Vision ### Text Processing - **Text Summarization Engine** - Implemented both extractive and abstractive methods - Utilized BERT, BART, and T5 transformers - Developed attention-based sequence-to-sequence models - **Advanced Text Generation** - Implemented GPT-3 integration - Developed Context Free Grammar systems - Created Markov Chain text generation - **Keyword Extraction System** - Implemented TextRank algorithms - Developed efficient extraction pipeline ### Data Generation & Processing - **GLOT500 Dataset Generation and Encryption** - Created dataset for low-resource languages - Implemented encryption mechanisms for data security - Focus on expanding linguistic diversity in NLP ### Computer Vision & Gaming - **Object Detection Systems** - Implemented Fast RCNN, RCNN, and YOLO - Developed detection for various object types - Created specialized tumor detection module - **Image Classification Platform** - Implemented CNN with data augmentation - Utilized transfer learning techniques - Processed CIFAR image dataset - **Pacman Game Implementation** - Developed pathfinding using BFS & A* algorithms - Created game interface using pygame - Implemented intelligent agent behavior ## Technical Skills Demonstrated - **Languages & Frameworks**: Python, JavaScript, HTML, CSS - **ML Frameworks**: TensorFlow, PyTorch, HuggingFace - **Web Frameworks**: Flask - **AI Models**: BERT, BART, T5, GPT-3, CNN, RNN, LSTM - *Computer Vision*: RCNN, YOLO, Transfer Learning - **Other Tools**: pygame, Djinga