# Data Quality Specialist

**Company**: Mistral
**Location**: France
**Work arrangement**: hybrid
**Experience**: mid
**Job type**: full-time
**Salary**: Competitive cash salary and equity
**Category**: Engineering
**Industry**: Technology
**Wikidata**: https://www.wikidata.org/wiki/Q22671858

**Apply**: https://jobs.lever.co/mistral/bd88179e-de69-4675-8a6c-74e2547a85ac
**Canonical**: https://yubhub.co/jobs/job_9e0a391d-70f

## Description

We're seeking highly motivated Data Quality Specialists with strong analytical skills and a keen eye for detail to join our Human Data Annotation team within the Science organisation.

This is a hybrid quality reviewing and tooling role: you'll spend the majority of your time reviewing and auditing code annotations against rubrics to ensure data used for training and evaluating AI models meets a high bar, and the remainder building, maintaining, and troubleshooting the internal tooling that annotators rely on day-to-day.

You'll collaborate closely with the annotators, technical program manager, and engineer stakeholders, and contribute to refining the guidelines and processes that shape how our data is produced.

### Key Responsibilities

- Generate and validate high-quality data annotations, based on guidelines and continuous feedback, for the development and evaluation of AI models

- Surface systemic issues, edge cases, and gaps in guidelines back to annotation operations and technical stakeholders

- Produce annotations yourself when needed, modeling the quality bar expected of the team

- Build and maintain internal tools and automation that streamline annotator workflows such as visualization dashboards, batch configuration scripts, output management utilities, and similar

- Troubleshoot environment, tooling, and CLI/git issues for annotators on their local machines, liaising with IT and engineering as needed

### About You

- A degree in computer science, engineering, or a related field. Alternatively, 2 to 5 years of professional experience in software engineering, technical support, or developing tools

- Hands-on experience using code agents (e.g. Mistral’s vibe) in your own development workflow, and genuine interest in how they're evolving

- Proficient in at least one programming language (e.g. Python, JavaScript, or similar), with enough breadth to read and reason about code across a few core languages

- Able to apply consistent judgment against a rubric and surface edge cases, ambiguities, or gaps in guidelines

- Sustained focus and accuracy on detail-oriented, high-volume review work

- Comfortable working in a Unix-like terminal: shell basics, package managers, environment setup, and git workflows (branches, merges, resolving conflicts)

- Able to troubleshoot local development environment issues (dependencies, virtual environments, paths, permissions) across common operating systems

- Professional proficiency in English, with strong writing and comprehension skills

### Nice to Have

- Prior experience in data annotation for AI/ML, especially LLM training (SFT, RLHF, preference data), evals/benchmarks, or agentic data

- Experience building an annotation team through interviews and training

- Experience supporting technical users or troubleshooting developer environments (internal tools support, DevRel, teaching assistant for coding courses, etc.)

- Fluency across multiple programming languages, or domain depth in one of: frontend, backend, DevOps, MLOps, data engineering

- Familiarity with rubric-based evaluation concepts, inter-annotator agreement, or quality measurement for human-labeled data

- Experience developing, deploying, and managing internal tooling or automation scripts

## Skills

### Required
- Python
- JavaScript
- Unix-like terminal
- Git
- Shell basics
- Package managers
- Environment setup