M
Ai 评测平台开发工程师（全栈）
Modelbest
北京
≈ 141,5 тыс.–235,8 тыс. ₽ · 1,8 тыс.–2,9 тыс. $ · 1,5 тыс.–2,6 тыс. €
12 000 – 20 000 CNY
🏢 Офис
Senior
Полная занятость
Продуктовая
Китай
Tech
Описание вакансии
Обязанности
•Build benchmark registry and evaluation dataset versioning
•Build evaluation result parsing and archival
•Create visualization dashboards
•Design AI evaluation platform architecture
•Develop evaluation analytics and metric aggregation
•Enable batch evaluation with retries and idempotency
•Generate evaluation reports and leaderboards
•Implement asynchronous evaluation task system
•Implement model promotion gatekeeping
•Integrate model checkpoint and inference services
•Integrate training inference and data platforms
•Manage judge model evaluation
•Orchestrate evaluation jobs
•Support version comparison and anomaly tracking
Технологии: Access Control, Asynchronous programming, Benchmark Registry, Cost Tracking, Data Modeling, Data Visualization, Database Design, Dataset versioning, Evaluation Dataset, Evaluation Dataset Versioning, Experiment Management, Failure diagnosis, Go, Idempotency, Java, LLM-as-a-Judge, Leaderboard, Logging, MLOps, Molecular Pipeline N/A, N A, Python, REST API, React, Result Analysis, Retry Mechanisms, Role-Based Access, Role-Based Access Control, Task Orchestration, Vue