Member of Technical Staff (Machine Learning Engineer, Ranking Quality - Search)

Навыки

• LLM
• Machine Learning
• Ответственность за результат
• Рекомендательные системы

О компании и продукте

• Perplexity is seeking an experienced Machine Learning Engineer to improve search quality across the middle and later stages of ranking. We are looking for a strong ranking generalist who can own ambiguous problems end to end and brings exceptional depth in either neural ranking or production ranking systems.

Задачи

• Relentlessly push search quality forward through models, data, evaluation, infrastructure, or any other leverage available
• Own ranking-quality problems end to end: define the evaluation, identify the bottleneck, build the solution, and ship it safely
• Train and evaluate retrieval, ranking, and classification models, including neural and LLM-based approaches where appropriate
• Build and operate ranking infrastructure, including feature computation, low-latency inference, multi-stage cascades, deployment, and monitoring
• Make sound trade-offs across quality, latency, reliability, cost, and engineering complexity
• Collaborate across Data, AI, Infrastructure, and Product while retaining ownership of the final quality outcome

Требования

• Deep understanding of search or recommender systems and their evaluation
• Proven ownership of a large-scale production ranking system or a substantial class of quality problems
• Strong machine-learning and software-engineering skills across data, models, serving, and monitoring
• Ability to drive ambiguous, cross-team problems without continuous task decomposition
• Exceptional depth in either modern neural ranking methods or low-latency ranking systems and runtime
• Minimum 5 years of relevant industry experience

Паспорт вакансии

История публикации

Откуда что взялось

Отмечено то, что вывели мы. Без пометки — значение назвал работодатель.

Почему на этом месте в выдаче

Порядок выдачи объявлен контрактом: свежесть решает между днями, полнота и зарплата — внутри дня.