GenAI Research Engineer: Distillation & Efficient Models

ByteDance; Seattle (WA); On-site; USD 242,000 - 456,000; Full time

Benefits offered by this job

Medical, dental, and vision insurance

401(k) with company match

Paid parental leave

Short-term and long-term disability

Life Insurance

Wellbeing benefits

10 paid holidays per year

10 paid sick days per year

17 days of Paid Personal Time

Job summary

ByteDance is seeking a Research Engineer/Scientist for the Vision-Applied Research team to design and implement efficient models for large-scale generative AI and multimodal understanding. You will work on distillation frameworks, model acceleration, and hardware-efficient inference to transfer capabilities from foundation models to smaller, deployable systems.

Responsibilities include developing efficient architectures and accelerating diffusion/autoregressive models, with a focus on practical

Qualifications

• B.S. in Computer Science or related field required or equivalent experience.
• Experience training generative AI or LLM models using PyTorch or JAX.
• Strong communication and collaboration in fast-paced teams.

Responsibilities

• Develop efficient algorithms and architectures for large-scale generative and multimodal models (distillation, quantization).
• Advance scalable generative modeling approaches focusing on diffusion and autoregressive methods for efficiency.

Skills

Efficient models

PyTorch

JAX

Strong communication

Collaboration

Education

B.S. in Computer Science or related field

Ph.D. in GenAI / MLSys (preferred)

Tools

PyTorch

JAX

Job description

ByteDance is seeking a Research Engineer/Scientist for the Vision-Applied Research team to design and implement efficient models for large-scale generative AI and multimodal understanding. You will work on distillation frameworks, model acceleration, and hardware-efficient inference to transfer capabilities from foundation models to smaller, deployable systems.

Responsibilities include developing efficient architectures and accelerating diffusion/autoregressive models, with a focus on practical