FamilyWall helps manage your Family's everyday life by sharing everyone's schedules and activities, tracking grocery lists, planning for dinner, managing to-dos as well as locating kids when they are outside. With FamilyWall, the whole family is on the same page!

We are seeking a highly motivated AI/ML Engineer - ideally a recent PhD graduate - to focus on the optimization and continuous improvement of AI models, particularly Large Language Models (LLMs). This role emphasizes model evaluation, fine-tuning techniques, dataset quality, and architecture experimentation to improve model performance.

You will work on high-impact AI initiatives that shape the future of user-facing features and internal intelligence tools, with a focus on rigorous evaluation, training iteration, and the pursuit of better architectures and outcomes.

Department: Engineering
Employment Type: Permanent - Full Time
Location: Paris
Workplace type: Hybrid

What you will accomplish:

Model Development & Optimization

Lead the full lifecycle of AI/ML models, focusing on performance improvements through fine-tuning, architectural experimentation, and continuous refinement.
Develop and maintain datasets for training and evaluation using real-world and synthetic generation methods.
Apply grounding strategies and fine-tune LLMs and vision models to enhance factual accuracy and reduce hallucinations.
Explore new modeling approaches and training methodologies that advance system capabilities.

LLM Evaluation & Continuous Learning

Design and maintain evaluation pipelines to systematically assess model quality across relevance, correctness, robustness, and user impact.
Conduct detailed error analysis and performance tracking to identify areas for continuous improvement.
Automate evaluation gates in CI/CD pipelines and support learning loops through strategic annotation and data sampling.

Tooling & Product Impact

Collaborate with Product and Data teams to align modeling goals with product outcomes and application needs.
Develop internal tools and utilities to support model experimentation, benchmarking, and validation processes.
Instrument applications for observability and help monitor model behavior in production environments.

Who you are:

A researcher-practitioner with a strong foundation in training, fine-tuning, and evaluating modern AI/ML models.
Focused on improving model performance through thoughtful experimentation, evaluation rigor, and high-quality data.
Collaborative, communicative, and motivated to solve real-world problems with cutting-edge techniques.
Capable of supporting engineering efforts where needed—especially those that enable better model evaluation and iteration.

What you bring:

PhD in Computer Science, Machine Learning, NLP, or a related field—recent graduates are strongly encouraged to apply.
Strong understanding of LLMs, fine-tuning methods (LoRA, PEFT, RLHF), and evaluation frameworks (LLM-as-a-judge, inter-annotator agreement, etc.).
Experience with Python and ML libraries such as PyTorch, HuggingFace Transformers, and vLLM.
Hands-on experience with dataset curation, training loop design, and architecture search.
Familiarity with tools such as Weights & Biases, MLflow, or LangChain for tracking and analysis.
Engineering experience (e.g., API exposure, model deployment) is a plus but not a strict requirement.

About FamilyWall

We believe technology can champion deeper connections within families, strengthen bonds, and improve communication. We're building solutions that foster connection, organization, and peace-of-mind throughout key stages and milestones of family life.

Apply Now