Perle is a data-curation and AI-evaluation platform founded by veterans from Scale, Amazon, Meta, and MIT. The hiring mix is heavily weighted toward data roles (58 active openings across annotation, curation, and quality assurance), with smaller engineering and research teams—a structure that reflects the labor-intensive nature of preparing training datasets at scale. Active projects span linguistic annotation, scientific reasoning evaluation, and adversarial feedback loops, pointing toward a platform designed for complex, human-guided model refinement rather than automated labeling alone.
Tech Stack 13 technologies
What Perle Is Building
◆Challenges
- Improving linguistic data quality
- Ensuring cultural precision
- Delivery delays
- Scaling accounting operations
- High-volume environment
- Ensuring cultural appropriateness
- Ensuring quality of annotated data
- High-volume workflow
- Ensuring accurate financial reporting
- High-quality training data
▲Active Projects
- Ai/nlp project
- Linguistic data annotation
- Evaluate advanced scientific reasoning produced by state-of-the-art ai systems
- Financial modeling and kpi reporting
- Pl/sql code example generation
- Training dataset annotation
- Ai language model training
- Ai voice data generation
- Scientific adjudication of ai outputs
- Provide research-grade corrective feedback
Hiring Activity
Steady95 roles · 40 in 30d
Department
Seniority
Company intelligence
