TL;DR
Software Engineer (AI): Developing and improving large scale pre-training multimodal AI models with an accent on reinforcement learning, data infrastructure, and model dynamics. Focus on designing and scaling multimodal vision models and advancing deep learning research.
Location: Mountain View, California, US
Salary: $141,000 - $202,000 per year
Company
hirify.global is a leading AI research company focused on advancing artificial intelligence for public benefit and scientific discovery with a strong emphasis on safety and ethics.
What you will do
- Conduct empirical and theoretical research to improve Gemini models and large scale pre-training.
- Work closely with the Multimodal Vision team to scale multimodal AI models.
- Focus on reinforcement learning and data infrastructure to enhance model dynamics.
- Analyze data and evaluations to support model training and improvements.
- Collaborate on deployment of large scale AI models.
Requirements
- Location: Must be based in Mountain View, California, US
- 2-5 years of professional software engineering experience.
- Strong knowledge of Python and/or C++ and experience with neural network training.
- In-depth understanding of Transformer models and LLM training dynamics.
- Ability to communicate technical ideas effectively.
Nice to have
- Experience with Large Language Models and Vision Language Models.
- Knowledge of JAX, PyTorch, or similar frameworks.
- Experience with data analysis, visualization, and ingestion pipelines for model training.
Culture & Benefits
- Work in a unique culture fostering long-term ambitious AI research.
- Collaborate with scientists, engineers, and machine learning experts.
- Commitment to diversity, equity, and inclusion.
- Competitive salary with bonus, equity, and benefits.
Hiring process
- Background check required upon offer.
- Recruiter will share specific salary details during hiring process.
Будьте осторожны: если вас просят войти в iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого - это мошенники. Обязательно жмите "Пожаловаться" или пишите в поддержку. Подробнее в гайде →