AI Engineer with less than a year in LLM Infrastructure and Production ML Systems
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
AI systems engineering student specializing in LLM infrastructure, retrieval-augmented generation (RAG), and production ML systems. Built end-to-end AI pipelines spanning evaluation frameworks, multimodal inference systems, and Dockerized MLOps workflows. Focused on evaluation-driven development, LLM observability, and scalable backend AI architectures for production deployment.
INSAT
Preparatory Cycle in Engineering
N/A – Present
Baccalaureate (Kairouan)
Scientific Baccalaureate (Honors)
N/A – Present
MUST University
Software Engineering
N/A – Present
MAT-IT
Web Development Intern (Drupal)
June 1, 2024 – August 31, 2024
India
Multi-Modal AI Health Assistant
June 1, 2026 – Present
Developed a multimodal ML inference system fusing medical image features and structured tabular patient data through a unified prediction pipeline. Integrated SHAP explainability layer for per-feature contribution attribution, enabling clinician-facing interpretability without model retraining overhead. Deployed production inference pipeline via FastAPI with Dockerized serving, input validation, async inference, and calibrated confidence scoring. Applied trustworthy AI design: out-of-distribution flagging, audit logging, and structured uncertainty quantification aligned with clinical deployment requirements.
Production ML Pipelines
June 1, 2026 – Present
Engineered DAG-based ML pipelines with Metaflow covering automated training, evaluation gating, artifact versioning, and full experiment tracking via MLflow. Containerized all pipeline stages with Docker for environment reproducibility; standardized re-entrant, monitored workflow execution for production ML lifecycle management.
Qanuny AI Legal Assistant (RAG + LLM Systems)
June 1, 2026 – Present
Architected and deployed a full-stack RAG system: document ingestion pipeline with recursive chunking, embedding generation, and FAISS / Chroma vector database integration. Optimized semantic retrieval with configurable similarity thresholds and hybrid search strategies, improving precision@k on domain-specific out-of-distribution legal queries. Built an LLM reasoning layer with structured prompt engineering for citation-grounded, context-aware legal response generation over a curated corpus. Exposed system via FastAPI REST backend with async inference endpoints, structured response schemas, and retrieval audit logging.
EvalGate - Evaluation-Driven LLM / RAG Framework
June 1, 2026 – Present
Engineered an evaluation-first LLM infrastructure framework for systematic debugging and optimization of RAG pipelines, replacing heuristic tuning with metric-driven iteration. Implemented automated LLM-as-judge evaluation suite measuring faithfulness, answer relevance, and cross-turn consistency at scale across retrieval pipeline variants. Designed iterative retrieval optimization loop: low-scoring traces automatically trigger targeted retrieval parameter adjustments and re-evaluation cycles. Built LLM observability tooling for prompt drift detection, context utilization analysis, and retrieval failure attribution — production-grade debugging infrastructure. Tracked all evaluation runs, metric deltas, and pipeline configs via MLflow for full reproducibility and experiment auditability.
View ProjectDigital Twin AI Platform
June 1, 2026 – Present
Built a simulation-based AI platform modeling real-world user and system behavioral patterns through iterative ML training and behavioral data replay pipelines. Designed full ML lifecycle: feature engineering → model training → evaluation gating → MLflow experiment registry with artifact versioning. Implemented Streamlit visualization layer for behavioral insight exploration, enabling interpretable outputs for both technical and non-technical stakeholders.
IBM Docker for DevOps
Unknown
June 1, 2026 – Present
IBM AI Engineering Professional Certificate
Unknown
June 1, 2026 – Present
Google Cloud MLOps Specialization
Unknown
June 1, 2026 – Present
AWS Machine Learning Fundamentals
Unknown
June 1, 2026 – Present
Cultural Fit Analysis
The candidate's project portfolio shows a strong interest and hands-on experience in cutting-edge AI domains like LLMs, RAG, and multimodal AI, which aligns well with an AI Engineer role. The diversity of projects (legal, health, digital twin) indicates adaptability and a broad technical curiosity. However, the lack of professional experience beyond a short internship might suggest a need for mentorship in a team environment.
Soft Skills & Operational Fit
The candidate's project descriptions indicate a proactive and structured approach to problem-solving, particularly in debugging and optimizing complex AI systems. The involvement in organizing events suggests organizational skills and teamwork potential. The focus on reproducible ML systems and observability aligns well with operational best practices in an AI engineering role.