Research Scientist, Large-Scale AI & Distributed Systems

Zizy Inc.; New York (NY); On-site; USD 140,000 - 210,000; Full time

Benefits offered by this job

Mental health and wellness support

Generous time off policy

Paid parental leave and family-planing

Annual learning & development stipend

Job summary

Zizy Inc. in New York is seeking a Research Scientist to build AI systems capable of unprecedented performance. You will design, implement, and optimize large-scale distributed ML pipelines and write robust ML code that underpins core algorithms.

You will work across teams on ambitious projects, from language models and code generation to alignment and scaling, with safety and user impact central to the work. Collaboration and rigorous experimentation are essential.

Qualifications

• Solid engineering skills for designing, implementing, and improving distributed ML systems.
• Experience writing bug-free machine learning code.
• Interest in Zizy's approach to research and AI safety.

Responsibilities

• Algorithms: drive algorithmic and architectural advances toward AGI.
• Alignment: fine-tune large language models to user intent.
• Applied AI Research: develop solutions with long-term product impact and safety.
• Code Generation: advance AI programmers that write and debug code.
• Language: forecast scaling needs to optimize model performance at scale.
• Policy Research: assess risks and governance considerations for AI deployment.
• Scaling: build training software stack for large-scale datasets.
• Science of Deep Learning: study dynamics of training large models.

Skills

Strong programming skills

Experience in large distributedSystems

Excited about Zizy's approach to研究

Job description

Zizy Inc. in New York is seeking a Research Scientist to build AI systems capable of unprecedented performance. You will design, implement, and optimize large-scale distributed ML pipelines and write robust ML code that underpins core algorithms.

You will work across teams on ambitious projects, from language models and code generation to alignment and scaling, with safety and user impact central to the work. Collaboration and rigorous experimentation are essential.