Чётко описанная роль уровня Staff в ведущей компании enterprise AI даёт опыт работы с продакшен-системами на базе LLM, распределённой инфраструктурой и внедрениями у клиентов; основные компромиссы - обязательная работа из офиса и широкий круг ответственности.
Кликните для подробной информации
Широкая ответственностьСовременный стекСильный AI-брендШесть недель отпускаВыше рынка
Оценка от Hirify AI
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
Staff Forward Deployed Engineer (AI): Architecting and deploying resilient large language model serving systems inside complex enterprise infrastructure with an accent on multi-tenant serving, multi-cluster failover, performance optimization, and custom hardware validation. Focus on benchmarking and profiling inference workloads, establishing production reliability and observability, and translating customer deployment challenges into platform improvements.
Location: United States; on-site. Base salary: $265,000–$325,000 in California, New York, and Washington; $240,000–$295,000 in other US states; C$340,000–C$415,000 in Canada, plus equity.
Company
Cohere is an enterprise AI company building foundation models and end-to-end products for real-world business applications.
What you will do
Embed with customer engineering teams to connect Cohere AI models to complex production environments.
Design multi-tenant model serving, intelligent routing, multi-cluster architectures, and cross-datacenter failover.
Optimize private and AI-in-a-box deployments across Kubernetes, SRE constraints, and model performance requirements.
Benchmark and profile customer LLM workloads, using metrics, dashboards, and distributed logs to diagnose performance degradation.
Support model deployment on GPUs, TPUs, AMD accelerators, and custom ASICs, and establish production SLOs and observability.
Translate customer deployment challenges into platform upgrades and core product improvements.
Requirements
8+ years of engineering experience running production infrastructure, designing large-scale distributed systems, or deploying enterprise software in complex environments.
Deep production Kubernetes expertise, including multi-cluster orchestration across hybrid and multi-cloud environments such as GCP, Azure, AWS, and OCI.
Strong understanding of LLM architecture and inference dynamics, including attention mechanisms, compute-bound prefill, and memory-bound decode phases.
Hands-on experience with inference engines and optimization stacks such as vLLM, TensorRT-LLM, lmd, or Dynamo.
Familiarity with distributed serving, prefill/decode disaggregation, distributed KV caching, and multi-GPU workloads.
Customer-facing technical leadership, including scoping ambiguous problems and guiding executive technical stakeholders.
Ability to write same-day production code while managing client deployment blockers and politics.
Strong programming foundations in Python, Go, or C++, with the ability to work across backend data pipelines and infrastructure tuning.
Culture & Benefits
Full health and dental coverage, mental health budget, RRSP matching, 401(k), and pension scheme.
100% parental leave top-up for up to six months for either parent.
Six weeks of paid vacation and annual enrichment benefits covering arts and culture, fitness, quality time, and workspace improvements.
Education and learning stipend for conferences, courses, and coaching.
Weekly lunch stipend of $75/£75 or local equivalent, plus office lunch programs, snacks, and community events.
$500 home office stipend, coworking benefit, office travel budget, and annual company offsite.
Hiring process
Applicants may be screened and assessed using AI-enabled tools against the position criteria.
Accommodation requests can be submitted through the provided accommodations form.
Будьте осторожны: если работодатель просит войти в их систему, используя iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого - это мошенники. Обязательно жмите "Пожаловаться" или пишите в поддержку. Подробнее в гайде →