Open to opportunities

Ahmed Mrabet

@ahmedmrabet

Message

AI/ML engineer focused on end-to-end production LLM systems, RAG, and agentic orchestration with measurable quality.

Morocco

Message

What I'm looking for

I’m looking to build and operate production LLM/AI systems end to end—FastAPI services, custom RAG/agents, cloud CI/CD, and strong eval/monitoring—while balancing quality, latency, and cost with real user feedback.

I build and operate production AI/ML systems end to end, turning LLM ideas into reliable services with FastAPI backends, custom RAG, and agentic orchestration using PydanticAI. I focus on measurable quality, not just demos—instrumenting and evaluating what matters.

In my current role as a Consultant Data Scientist, I designed cloud architecture and shipped production GenAI services with CI/CD, deployment, and monitoring. I built a production knowledge-management assistant using a SharePoint ingestion pipeline plus a continuous feedback loop for evaluation and quality improvement.

I also delivered high-impact systems: an OCR + anomaly detection + structured extraction service processing 20k+ contracts/year and protecting 120M MAD exposure through signature and stamp validation. I built an insurance-quote comparison application integrating Document Intelligence with LLM reasoning, and a customer-conversation mining system that generates an enhanced support FAQ with a 93%-accurate request-classification model routing requests in production.

My engineering approach is deeply first-principles, including implementing GPT behavior from scratch and exploring agentic architectures for practical LLM workflows. I’m comfortable trading off output quality, latency, and cost while owning cloud infrastructure, using Azure + Terraform and aligning services with OpenTelemetry for monitoring and evaluation.

Experience

Work history, roles, and key accomplishments

Current

Consultant Data Scientist

Current

Artefact

May 2024 - Present (2 years 1 month)

Built an end-to-end production knowledge-management GenAI assistant, designing Terraform-based cloud architecture and deploying an async FastAPI backend with CI/CD plus a custom RAG framework and continuous quality evaluation. Shipped OCR + anomaly detection for 20k+ contracts/year and protected 120M MAD via signature/stamp validation, and developed production quote-comparison and support-Faq auto

Terraform fastAPI RAG CI CD LLM Evaluation OpenTelemetry SQL Azure

Data Scientist Intern

Artefact

Feb 2024 - May 2024 (3 months)

Compared agentic architectures (ReAct, planner-executor, multi-agent) for a natural-language-to-SQL interface and benchmarked QLoRA fine-tuning on StarCoder and Llama 3, tuning for accuracy vs. cost. Delivered a working, commercially oriented demo evaluated for fit.

Agent Architectures React StarCoder LLaMA Hyperparameter Tuning Benchmarking

Machine Learning Engineer Intern

Yakeey

Jun 2023 - Aug 2023 (2 months)

Built an end-to-end satellite imagery pipeline to extract building footprints, benchmarking deep-learning segmentation models and designing preprocessing/post-processing. Deployed the solution behind a FastAPI service in Docker.

Satellite Imagery fastAPI Docker Benchmarking Model Serving Computer Vision

Software Engineer Intern

KubicBits

Jun 2022 - Sep 2022 (3 months)

Engineered a vehicle-routing optimization algorithm, improving runtime by 30% and solution quality by 23%, and deployed it to production. Built the backend using Django REST Framework with PostgreSQL/PostGIS.

Django REST Framework PostgreSQL PostGIS REST APIs Geospatial Data

Education

Degrees, certifications, and relevant coursework

ENSIAS

Diplôme d’Ingénieur d’Etat, Artificial Intelligence

2021 - 2024

Activities and societies: Specialization: machine learning, deep learning, NLP, computer vision, distributed systems.

Earned a Diplôme d’Ingénieur d’État (M.Eng) specializing in Artificial Intelligence. Coursework covered machine learning, deep learning, NLP, computer vision, and distributed systems.