ML Engineer: Production AI for Legal Tech (Zurich)

DeepJudge AG; Zürich; Hybrid; CHF 90.000 - 130.000; Vollzeit

Benefits dieser Stelle

Office near Zurich central station

Ownership from day one

Zusammenfassung

DeepJudge is seeking a Machine Learning Engineer to design, build, evaluate, and refine AI systems powering our core products. You will own the entire lifecycle of an ML-enabled component—from research to production—including data analysis, modeling, implementation, delivery to customers, and offline/online evaluation.

Location: our Zurich-area office; you will work mostly in the office from day one, shaping high-impact ML solutions with a fast-paced, cross-functional team.

Qualifikationen

• Strong hands-on experience training and serving deep learning models.
• Proficient in PyTorch, JAX or TensorFlow.
• Experience building scalable AI-powered systems delivering real value.
• Familiar with neural retrieval or agentic systems is a strong plus.
• Take ownership of key product components and communicate technical vision.
• Thrives in fast-paced environments and collaborates across teams.
• Excellent communication skills and ability to share complex ideas.

Aufgaben

• Bridge Research and Production: translate research into scalable, production-ready systems.
• Core ML: train, evaluate, and optimize ML systems for robust inference and deployment.
• Search: advance search quality using models, data, tools, and evaluation methods.

Kenntnisse

Deep learning

PyTorch

TensorFlow

JAX

Model deployment

Search/IR

Neural retrieval

Python

Jobbeschreibung

DeepJudge is seeking a Machine Learning Engineer to design, build, evaluate, and refine AI systems powering our core products. You will own the entire lifecycle of an ML-enabled component—from research to production—including data analysis, modeling, implementation, delivery to customers, and offline/online evaluation.

Location: our Zurich-area office; you will work mostly in the office from day one, shaping high-impact ML solutions with a fast-paced, cross-functional team.