About
I’m Mathieu Ajaka.
I’m an AI engineer and solutions architect. I work with organizations in Saudi Arabia, the UAE, and Europe to turn operational and data problems into production AI systems: scoping requirements, designing the architecture, building with Python and FastAPI, and proving behavior with structured evaluation. I build bilingual Arabic–English systems as a specialty.
Based in / serving
Saudi Arabia · UAE · Europe
Languages
Arabic (native) · French (native) · English (professional)
Open to AI Solutions Architect, AI Engineer, GenAI, and Applied AI roles (with a current focus on Saudi Arabia) as well as consulting engagements. (Work-eligibility details available on request.)
Experience
Experience first, because that’s what the target roles are hired on. Before the titles: I’ve scoped requirements with stakeholders, written the architecture, presented it to technical and non-technical audiences, built the systems, and stayed accountable after deployment.
-
AI Solutions Architect · WAJEEH Agency
- Sole technical hire and de facto technical lead at an AI agency serving Saudi enterprise clients across CX, brand, and regulatory domains, owning solution design from scoping through production operations.
- Designed AI solutions for ~12 Saudi enterprise accounts: LangGraph-orchestrated agentic workflows with state machines, human-in-the-loop checkpoints, and tool-use routing across Arabic and English pipelines.
- Built a model-routing layer (local Llama 3.1 8B / Mistral 7B for templated content; Claude Sonnet / GPT-4o for regulated and executive-facing outputs) and an evaluation discipline (RAGAS, LLM-as-judge, QLoRA adapter release gates) with a CI/CD golden-set gate on every change.
-
AI Engineer (contract) · ARA Technologies MEA
- Diagnosed and rebuilt the ML and data platform behind an analytics product; provided technical direction to two internal engineers.
- Resolved an Elasticsearch p95 regression (7 nodes, ~11M docs) and re-architected cron ingestion to Kafka with at-least-once delivery and idempotent upserts; rebuilt client segmentation with XGBoost on time-based holdout, introducing MLflow as the model-lifecycle backbone.
-
Freelance AI Solutions Engineer · Independent
- Three engagements as sole technical owner: bilingual legal document intelligence for a Riyadh workforce-compliance entity (domain-aware chunking, multilingual-e5-large + BM25/RRF hybrid retrieval, deployed to an air-gapped environment); ingestion and anomaly detection for an Abu Dhabi analytics provider; and retrieval + segmentation for a French e-commerce platform.
-
Data & ML Analyst / Data Analyst (project-based) · BREI (Paris) · ALUSTEEL Group (Lebanon)
- First production ML work: Python/SQL telemetry pipelines and Isolation Forest anomaly detection with threshold calibration; earlier, consolidated operational data into a unified analytics layer and automated weekly reporting.
Education
- MSc, Applied Artificial Intelligence: Université Saint-Joseph (USJ), Beirut, 2024–2026. Thesis: TCN vs. ANFIS under concept drift on financial time-series (BTC-USD and ETH-USD, 2018–2024).
- BSc, AI and Business: IA School, Paris, 2021–2024.
Certifications
- Deep Learning Specialization (Andrew Ng), Coursera, 2023.
- Google Cloud Professional ML Engineer, in progress, expected 2026.
Working together
The case studies show how I approach systems; the CV has the full timeline. If you’re hiring or have a project, the fastest way to start is a short note about what you’re trying to do.