Data Scientist - GenAI Engineering

Location: Casablanca

Type: Full-time

Experience Level: Mid–Senior

About the Role

We’re looking for a Data Scientist specialized in Generative AI and Prompt Engineering to design, fine-tune, and deploy intelligent agentic systems across our AI product ecosystem. You’ll work on cutting-edge projects leveraging models such as GPT-5, DeepSeek, Claude, Gemini, and Mistral, integrating them with LangChain, n8n, and custom orchestration flows to create scalable multi-agent applications.

Key Responsibilities
* Design, fine-tune, and evaluate LLM-based models for real-world tasks (content generation, reasoning, classification, recommendation, etc.)
* Develop and optimize prompt architectures, function-calling schemas, and tool-augmented reasoning flows.
* Build agentic AI pipelines using LangChain, n8n, or similar frameworks for automation, retrieval-augmented generation (RAG), and task orchestration.
* Integrate external APIs and data sources (search, CRM, analytics) into AI workflows.
* Collaborate with engineers to productionize and monitor GenAI services (FastAPI, Docker, vector databases, etc.).
* Research and benchmark new foundation models (e.g., DeepSeek, Claude, Gemini, LLaMA 3, Mistral).
* Contribute to continuous improvement of model performance, latency, and cost optimization.

Requirements
* Strong background in Python, machine learning, and LLM APIs (OpenAI, Anthropic, DeepSeek, Google, etc.)
* Experience with LangChain, LlamaIndex, or n8n for building agentic flows.
* Solid understanding of prompt engineering, embeddings, RAG pipelines, and model evaluation metrics.
* Hands-on with FastAPI, Pydantic, Docker, and vector databases (Pinecone, Chroma, Weaviate, etc.)
* Familiarity with data analysis and NLP (transformers, fine-tuning, token optimization).
* Excellent problem-solving, research, and communication skills.

Nice to Have
* Experience training or fine-tuning LLMs or multimodal models.
* Familiarity with workflow automation (Zapier, Make, n8n) and AI agents orchestration.
* Exposure to real estate, e-commerce, or lead intelligence domains.