Data Science and Artificial Intelligence Engineering Student
Applied AI | Machine Learning | Deep Learning | Time Series | NLP | Computer Vision | FastAPI | Full-Stack AI Systems
I build applied AI systems that connect rigorous modeling, clean software architecture, and usable product interfaces. My work is strongest where research has to become something concrete: an API, a dashboard, a decision-support tool, or an AI assistant that people can actually use.
- Research discipline: I care about validation, leakage control, benchmarks, and reproducible experiments.
- Engineering structure: I separate research code, production services, APIs, UI, tests, and documentation.
- Product thinking: I do not stop at notebooks; I like turning models into clear tools and workflows.
- AI system design: I am interested in RAG, local LLMs, controlled generation, guardrails, and human-facing AI interfaces.
- Learning mindset: I enjoy complex projects that force me to connect statistics, ML, software engineering, and design.
My main areas of interest include:
- Machine Learning and Deep Learning
- Natural Language Processing
- Computer Vision
- Time-Series Forecasting
- Statistical Modeling and Econometrics
- Financial Risk Analytics
- Backend and Frontend Development
- UI/UX-oriented AI Applications
- FastAPI, Java, Python, and full-stack software engineering
- RAG systems, local LLMs, prompt engineering, and AI guardrails
End-to-end market risk forecasting and AI-assisted risk analysis platform, applied to the Moroccan MASI stock index.
It combines ML forecasting pipelines, VaR/Expected Shortfall backtesting, HMM volatility regimes, a FastAPI dashboard, trained artifacts, and a controlled local RAG chatbot.
Research companion for the market risk dashboard.
The notebooks cover statistical validation, GARCH/EGARCH benchmarks, LSTM VaR/ES modeling, HMM regime analysis, backtesting, and economic evaluation.
Reference architecture and implementation for a controlled RAG chatbot designed for market risk dashboards.
It includes embedding-based intent routing, Chroma vector retrieval, local LLM integration, response policies, and guardrails against hallucinated metrics or financial advice.
Library management API built with FastAPI and Clean Architecture.
It includes JWT authentication, MySQL persistence, reservations, borrowings, favorites, reviews, and structured backend layers.
Frontend prototype for a DFC/ECO workflow dashboard prepared for a competition submission.
The project focuses on operational UI, workflow screens, dashboard layout, and clean interaction structure.
I am currently focused on building stronger applied AI systems around:
- time-series forecasting and risk analytics;
- local RAG assistants for technical dashboards;
- backend APIs for ML-powered products;
- clean project documentation and reproducible workflows.
- GitHub: mohamedzayd-elfahime
- LinkedIn: Mohamed Zayd Elfahime