Mohamad Nesr

AI Back‑Documentation • LLMs: Mixtral 7x8B Instruct, GPT‑3.5/4, Llama 2, Qwen • APIs: OpenAI, Ollama, vLLM • Few‑shot learning and prompt engineering • Tool/function calling and multi‑agent setups • n8n workflows with a dedicated AI agent • Benchmarking with clear performance reports • Agent development in Java (LangChain4j) • Python data collection and processing • Git RAG • Embeddings, tokenization, retrieval + generation • LlamaIndex and Haystack • Hybrid approach: SentenceTransformers + BM25 • Multi‑source extraction and indexing Fine‑tuning • Transformers, TRL, PyTorch • Target: microsoft/phi‑2 • PEFT (LoRA), quantization • BLEU, ROUGE • Data cleaning, preprocessing, structuring • Final indexing

This internship's objective is creating a chatbot to assist users while solving applicative crises. • Product: a decision-support AI chatbot for applications incidents resolution • Applied NLP: classification, intent extraction, automated responses • Data: ingestion, cleaning, transformation, and quality monitoring • Cloud stack: AWS (Lex, Lambda, S3, Kendra)

During my internship, I developed a C++ microservice that integrates a real time object detection system called Yolov4. This internship allowed me to gain many skills such as: • C++ microservice developement • Software architecture • Image processing (OpenCV) • AI (machine learning + data training & testing) • Agile/Scrum methodologies • Git