Agentic Harness Engineer

DeepJudge seeks a Senior Agentic Harness Engineer in Zürich to build and maintain the core system that turns AI models into reliable enterprise applications, ensuring high agent quality and driving product decisions.

This position serves as the central pillar for DeepJudge's product, responsible for converting advanced AI models into dependable, enterprise-ready applications. The engineer will ensure the quality of AI agents across various use cases, playing a mission-critical role where technical choices directly shape the customer experience.

Responsibilities

• Manage the agentic harness that supports key features within the DeepJudge platform.
• Design, implement, and oversee end-to-end evaluation processes, including setup, instrumentation, and ongoing monitoring.
• Enhance agent performance by modifying harness code and refining prompts.
• Provide input on model selection and product launches based on agent quality metrics.
• Collaborate with Legal Engineering, Product, and Customer Success teams to prioritize and implement harness enhancements.
• Monitor emerging trends in LLMs, GenAI, and agentic AI to integrate best practices into the system.

Requirements

• Master's degree in Computer Science or a related discipline, or equivalent hands-on experience.
• Proficiency in backend programming languages such as Python, Rust, Go, or C++.
• Expertise in evaluating agentic systems and creating effective evaluation frameworks.
• Strong enthusiasm for current developments in GenAI and frontier models.
• Capability to develop solutions independently from the ground up.
• Proven track record of building large-scale applications powered by LLMs.

Nice to have

• Prior experience working with agentic SDKs.
• Familiarity with evaluating LLM-based systems.

What the company offers

• Significant ownership and the ability to influence the company, product, and culture from the start.
• Opportunity to work in one of the most dynamic sectors in AI.
• Office located just a five-minute walk from Zurich central station, with a primarily onsite work model.
• Supportive and collaborative team environment.
• Opportunities for professional growth and career advancement within the organization.

About the company

DeepJudge was established by former Google search engineers holding PhDs in AI from ETH Zurich. The company provides world-class enterprise search and AI infrastructure that enables legal teams to leverage their internal knowledge. Its platform is trusted by leading law firms and legal departments globally. The organization values urgency, rigorous thinking, and close client collaboration to solve meaningful problems, while fostering an inclusive, high-performance culture built on diverse perspectives.

• Founded by former Google search engineers with PhDs in AI from ETH Zurich.
• Helps legal teams unlock and apply the knowledge inside their organizations through world-class enterprise search and AI infrastructure.
• Platform is trusted by many of the world’s leading law firms and legal teams.
• We move with urgency, think rigorously, and work closely with our clients to build products that solve meaningful problems.
• We believe great teams are built on diverse perspectives and experiences.
• Committed to fostering an inclusive, high-performance culture where everyone can thrive