Staff Machine Learning Engineer (AI Agents)
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Описание вакансии
TL;DR
Staff Machine Learning Engineer (AI Agents): Building and optimizing production-grade autonomous AI agents for customer service with an accent on iterative planning architectures, tool-use, and memory management. Focus on designing domain-specialized models via reinforcement learning and implementing enterprise-scale guardrails for multi-agent systems.
Location: Must be based in Australia (Remote, Melbourne, or Sydney)
Company
provides customer experience software that powers billions of interactions, helping businesses build meaningful relationships with their customers.
What you will do
- Advance the core iterative architecture of autonomous AI agents for complex ticket resolution.
- Develop RL training infrastructure to create domain-specialized models for customer service.
- Design and implement multi-layered guardrails to prevent tool misuse and hallucinations at scale.
- Build robust evaluation suites and CI-integrated quality gates for agent performance.
- Collaborate on multi-agent delegation patterns and memory management for concurrent sessions.
Requirements
- 5+ years of experience building production ML/AI systems.
- Hands-on experience with agent architectures (planning, tool dispatch, memory, error recovery).
- Fluency in Python and PyTorch.
- Strong evaluation instincts with experience building internal benchmarks.
- Must be based in Australia.
Nice to have
- Depth in Reinforcement Learning (RL) for language models (reward shaping, online/offline tradeoffs).
- Experience with agent frameworks and judgment on custom implementation.
Culture & Benefits
- Hybrid work model allowing for flexibility and in-person collaboration.
- Inclusive and diverse global work environment.
- Focus on meaningful work and reducing complexity in customer service.
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 →