Software Engineer (Foundational AI)

Maisa AI is seeking a Foundational AI Engineer to join our growing team and work directly with our leadership on building proprietary models. This role focuses on training, fine-tuning, and evaluating small and large language models that power Maisa Studio, enabling our clients to reduce dependency on frontier providers while maintaining high performance in regulated environments.

Key responsibilities

• Train, fine-tune, and post-train proprietary language models from data preparation through to release.
• Design and execute rigorous model evaluation strategies, including benchmarks and quality criteria tailored to real-world client use cases.
• Develop and optimize training pipelines, including data curation, experiment tracking, and infrastructure reproducibility.
• Collaborate with AI infrastructure teams on model serving and inference optimization to ensure reliable production deployment.
• Monitor deployed models for quality degradation and drive necessary retraining or corrective actions.

Requirements

• Proven, hands-on experience in training and fine-tuning machine learning models.
• Strong proficiency in PyTorch with the ability to build and modify model architectures directly.
• Solid software engineering fundamentals in Python.
• A rigorous approach to model evaluation and a deep understanding of how to detect regressions.
• Ability to thrive in a fast-moving, early-stage environment with a high degree of ownership.

What we offer

• The opportunity to work at one of Europe’s most exciting early-stage AI startups on cutting-edge Agentic Process Automation.
• Direct influence on how major multinational organizations transform their critical business processes.
• A collaborative culture where you have a voice in shaping how we work and build together.
• A fully remote working environment with a team of deep technical and industry experts.