Project Perseus | Data Labeling Associate

About the role

Welo Data is actively searching for meticulous Data Labeling Associates who demonstrate genuine curiosity, precision in execution, and the capacity for independent critical thought. This role is distinctly different from standard content moderation positions because it focuses on evaluating advanced AI system behavior rather than simply filtering user content. You will work directly alongside sophisticated AI systems, carefully assessing their results, uncovering hidden limitations, and contributing insights that lead to improved real-world behavior. Success in this position hinges on a dual capability of rigorously adhering to established processes while simultaneously offering insightful analysis that pushes the project forward. You are expected to engage in technical discussions with peers, team leads, and project stakeholders in a clear and professional manner. Ultimately, your work will translate complex observations into concrete recommendations that drive tangible model improvements and shape future development strategies.

Key facts

What you'll do

Design intake procedures that capture unusual edge cases and ambiguous prompts for thorough and systematic review.

Construct organized guidelines that convert high-level instructions into specific, repeatable rules that the entire team can follow consistently.

Evaluate model outputs based on established criteria, highlighting patterns that suggest potential risk, unreliability, or misalignment.

Produce evaluation summaries that clearly explain why certain responses are insufficient and propose minimal, actionable corrections.

Collaborate with engineering professionals to ensure that annotation practices remain aligned with training and deployment methodologies.

Refine review checklists dynamically so that each labeling session addresses the most essential and relevant behaviors for the current project phase.

Work with team leads to reprioritize tasks efficiently when new risks emerge in model outputs or when business objectives shift unexpectedly.

Strictly observe all location requirements, ensuring that all work is performed exclusively within an approved city and adhering to onsite policies.

Conduct detailed comparative analysis between multiple model responses to identify subtle differences in reasoning and factual accuracy.

Identify systematic gaps in model performance by aggregating labeling findings and summarizing trends for broader review.

Maintain a high level of consistency across long work sessions by applying structured techniques and rigorous self-checking methods.

Document nuanced labeling decisions with clarity, providing justifications that support future audits and process refinement.

Support the creation of training materials derived from labeling insights that help onboard new team members effectively.

Proactively flag ambiguous scenarios or novel failure modes so that they can be addressed before scaling the evaluation effort.

Requirements

You must have authorization to work in the United States without sponsorship, ensuring that you can begin employment immediately under the required terms.

You must be available to work 40 hours per week over a standard five-day schedule without exception.

You must commit to a one-year contract term, with the possibility of extension contingent upon performance and project needs.

You must live in or be able to travel regularly to one of the designated on-site cities listed in the job details.

You must be capable of engaging in technical discussions with engineering teams and cross-functional stakeholders.

You must demonstrate a strong attention to detail while maintaining awareness of the broader project goals and system-level implications.

You must be comfortable working independently when guidance is limited, making sound judgments based on established criteria and context.

You must adhere strictly to all location constraints, ensuring that work is never performed outside approved areas.

You must be willing to adapt to changing priorities as project requirements evolve and new risks are identified.

You must be able to follow complex instructions and apply them consistently across diverse evaluation tasks.

You must possess the legal right to work in the United States for the duration of the contract without requiring visa sponsorship.

You must be prepared to commit to the full duration of the contract as outlined in the hiring terms.

Practical notes

The role operates on a full-time schedule, requiring a standard 40-hour workweek spread across five days, ensuring consistent availability for collaborative work and synchronous discussions. All work must be conducted onsite in one of the approved locations, and remote work arrangements are not permitted under any circumstances. Candidates must be prepared to relocate or commute to a designated city to meet this requirement. The position is structured as a 1-year contract with a defined possibility of extension, subject to performance reviews and evolving project demands. No specific software tools or prior experience in similar domains are explicitly required, allowing the role to remain accessible to a diverse range of qualified candidates. Please review the official application page for any updates to hours, travel expectations, or deadlines that may affect your eligibility.