Чёткая и значимая роль в быстрорастущей компании по AI-инфраструктуре: сильная компенсация, доля в компании и практическая работа с перспективной областью LLM-систем.
Актуальные технологии
Большая ответственность
Перспективная сфера
Сильный пакет
?Зарплата на уровне рынка
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
Software Engineer (AI Inference) (HPC/LLM): Building benchmarking, profiling, observability, and development tools for high-performance AI inference infrastructure with an accent on GPU systems, model evaluation, and runtime performance. Focus on automating performance testing, identifying latency-cost-quality trade-offs, and optimizing model serving across distributed compute and networking stacks.
Location: Hybrid in San Francisco, Montreal, New York, Seattle, or Toronto
Salary: $165,000–$330,000 per year, plus equity
Company
Baseten provides AI inference infrastructure, applied AI research, and developer tooling for companies bringing machine-learning models into production.
What you will do
Evaluate and automate LLM quality benchmarks and custom performance suites for long-context, KV-cache reuse, and disaggregated serving workloads.
Develop GPU-enabled internal development environments for high-performance model experimentation.
Build and contribute to open-source benchmarking and model-evaluation tools, including InferenceMAX and genai-bench.
Profile systems with PyTorch Profiler, NVIDIA Nsight Systems, and py-spy to identify compute and networking bottlenecks.
Create real-time dashboards, alerts, CI/CD performance tests, and release automation for model runtimes.
Develop optimization tools to identify the best latency, cost, and quality configuration for each model and workload.
Requirements
Mid-to-senior engineering experience with strong technical depth and communication skills.
Understanding of GPU memory subsystems, InfiniBand, and data movement across clusters.
Experience or strong interest in scripting, stress testing, and fuzz testing to identify system limits.
Curiosity about Transformer mathematics, FLOPs, and memory requirements.
Familiarity with Python and willingness to master the NVIDIA software stack.
Ability to drive cross-team initiatives, work through ambiguous requirements, and mentor engineers.
Nice to have
Familiarity with C++.
Culture & Benefits
High ownership and autonomy while leading a small team and building tools from scratch.
Opportunities to contribute to open-source projects and develop expertise in GPU orchestration and LLM inference.
Meaningful equity and competitive compensation.
Flexible PTO, including a company-wide winter break, and paid parental leave.
U.S. employees receive full medical, dental, and vision coverage, plus access to a company-facilitated 401(k).
Fertility and family-building support through Carrot.
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 →