Software Engineer, Evaluation Frontend

Responsibilities

• Build and maintain frontend tools and dashboards to visualize and analyse large-scale model evaluation data stored in central databases.
• Design and implement statistical and visualization components that help researchers and engineers compare hundreds of model evaluations and extract signal from noisy results.
• Collaborate closely with cross-functional teams (researchers, backend engineers, data scientists) to define UX and data requirements for evaluation tooling.
• Implement performant, accessible, and testable UIs using modern frontend frameworks; ensure smooth integration with backend APIs and data pipelines.
• Own instrumentation, performance monitoring, and iterative improvements based on user feedback from modelling teams.

Requirements

• Frontend Development
• Statistical Analysis
• Data Visualization
• Python Proficiency
• ML Frameworks
• Experiment Design
• API Integration
• Strong Communication

Preferred Qualifications

• Experience building frontend visualization systems and dashboards for research or ML teams.
• Familiarity with ML evaluation methodologies and experiment analysis for language models.
• Proficiency with modern frontend stacks (React, TypeScript) and visualization libraries.
• Experience with backend integrations, data pipelines, and working with large datasets.
• Prior publications or presentations in top-tier ML venues (NeurIPS, ICML, ICLR, MLSys, ACL, EMNLP) are a strong plus.

Benefits

• Competitive health and medical coverage
• Generous annual and paid leave
• Remote-flexible / hybrid working
• Stipends for co-working, lunches, and personal enrichment
• Parental leave top-up and wellbeing support

About the Company

Cohere builds and deploys frontier large language models and developer tools to scale intelligence for customers worldwide. We combine research and engineering to deliver models and products used for content generation, semantic search, retrieval-augmented generation, and agent systems. Cohere values diversity and inclusivity and seeks candidates who are passionate about building tools that enable scientific decisions and improve model development workflows. This role is based in Dubai (UAE) with hybrid and remote options within the UAE.