I'm a technical problem-solver working at the intersection of AI engineering and client delivery. I build systems that have to hold up in production, not just in a notebook, and I'm most energized when I'm close to the problem: understanding what a business actually needs, architecting a solution, and staying hands-on through implementation until it's delivering real value.

My background is in Generative AI, NLP, RAG, and knowledge graphs, but the throughline across my work is deployment: taking a technical capability and making it fit into a real workflow, for a real user, under real constraints.

Areas of Focus:

• Solution Architecture & Technical Consulting
• Client-Facing Implementation & Deployment
• Generative AI & Large Language Models
• Retrieval-Augmented Generation (RAG) & GraphRAG
• AI Agents & Tool Integration
• Full-Stack & Backend Systems
• Data Engineering & Analytics
• Cloud & AI Infrastructure

AI Developer Intern | Axe Finance AI for Financial Performance Analysis | February – June 2026 Deployed a production-grade GraphRAG platform for financial risk analysis, working directly with stakeholders to translate financial-analysis requirements into a system combining financial reports and external peer data in a Neo4j knowledge graph. Designed multi-step reasoning and peer-benchmarking pipelines, achieving a 34% improvement in retrieval precision, 92.3% faithfulness, and 84.8% answer relevancy through RAGAS-based evaluation. Shipped as a containerized FastAPI/Docker REST API with an independent LLM-judge evaluation layer so output quality could be monitored in production. Technologies: Neo4j, LangChain, GraphRAG, RAG, FastAPI, Docker, RAGAS, LLMs

AI Developer Intern | Olindias June – August 2025 Scoped and delivered an end-to-end semantic search pipeline: scraped and structured source data with BeautifulSoup and Playwright, applied RAG with MPNet embeddings for retrieval, and exposed the system through a FastAPI service ready for integration. Technologies: Python, Playwright, BeautifulSoup, RAG, LangChain, Sentence Transformers, FastAPI

AI Software Developer Intern | DabbaDabba

Built and deployed a multilingual OCR and Transformer-based pipeline for document translation and structured data extraction using Tesseract, mBART, and Hugging Face models, reaching 95.7% field-extraction accuracy across 50 languages in a real document-processing workflow. Technologies: Python, OCR, Tesseract, Transformers, mBART, Hugging Face, NLP

Full-Stack Developer Intern | GPRO Consulting July – August 2024 Delivered enterprise web applications for consulting clients using Spring Boot and Angular, building REST APIs and backend services with a focus on secure, scalable, maintainable software that met client requirements from day one. Technologies: Spring Boot, Angular, REST APIs

Built and deployed a knowledge-graph-based RAG system for extracting financial information and surfacing risk signals. Combined LLM-based information extraction, vector embeddings, Neo4j knowledge graphs, and Amazon Bedrock to enable contextual retrieval and reasoning over financial data — architected with production integration in mind. Technologies: Python, LangChain, Neo4j, GraphRAG, Amazon Bedrock, Qwen3, Embeddings

Clustered 277,000+ search queries using Sentence-BERT embeddings and K-Means, with UMAP dimensionality reduction and an interactive Streamlit visualization for stakeholder review. Technologies: Python, NLP, Sentence-BERT, K-Means, UMAP, Streamlit

Built an LLM-powered calendar assistant that understands natural-language scheduling requests and integrates with Google Calendar through an API-based architecture designed for real-world use. Technologies: FastAPI, Angular, LLMs, Ollama, Llama 3, API Integration

Implemented a one-shot facial verification system using a Siamese neural network architecture to learn similarity between facial images. Technologies: TensorFlow, Deep Learning, Computer Vision, Siamese Networks

Developed a deep learning system combining generative and convolutional models for signature generation and forgery detection. Technologies: PyTorch, GANs, CNNs, Deep Learning, Computer Vision

Delivered a full-stack sustainable ecotourism platform connecting users with eco-friendly destinations, activities, reviews, and travel experiences in Tunisia, from requirements through deployment. Technologies: React, Flask, REST APIs, MySQL, JWT

Client-Facing & Delivery Solution Architecture, Technical Consulting, Requirements Gathering & Translation, Stakeholder Management, Rapid Prototyping, Client Implementation & Onboarding, Cross-functional Collaboration

AI & Machine Learning Python, PyTorch, TensorFlow, Keras, scikit-learn, Hugging Face Transformers, Sentence Transformers, NLTK

Generative AI & LLMs RAG, GraphRAG, LangChain, LLM Integration, Prompt Engineering, Embeddings, Vector Search, Knowledge Graphs, Ollama, Amazon Bedrock

Backend Development FastAPI, Flask, Spring Boot, REST APIs, SQLAlchemy, JWT

Frontend Development React, Next.js, Angular, HTML, CSS, JavaScript

Data & Databases Pandas, NumPy, SQL, MySQL, PostgreSQL, MongoDB, ChromaDB, Neo4j

DevOps & Tools Docker, Git, GitHub Actions, Linux, CI/CD

Bachelor's Degree in Business Administration — Major in Information Technology, Minor in Business Analytics Tunis Business School

LinkedIn: linkedin.com/in/saja-moussa Portfolio: sajaa45.github.io/portfolio Email: moussa.saja@gmail.com

"Turning complex technology into solutions that actually work for the people using them."