ByteDance is looking to develop and operate a big model service platform that offers businesses Model-as-a-Service (MaaS) solutions. This involves researching and developing advanced techniques for LLMs, including model post-training, fine-tuning, reinforcement learning, reasoning, evaluation, prompt optimization, multi-modality, and LLM application/agent development.
Requirements
- Have prior experience working with training or inference of large language models (PyTorch or Tensorflow).
- Strong understanding of cutting-edge LLM research (e.g., long context, multi modality, finetuning, alignment, RL, agent, etc.) and possess practical expertise in effectively implementing these advanced systems.
- Proficiency in programming languages such as Python or C++ and a track record of working with deep learning frameworks or agent frameworks.
- Experience with building LLM Applications, have a deep understanding of Agentic frameworks, RAG, Test-time scaling, Reasoning, Evaluation, Multi-modality, etc.
- Experience with deploying AI applications into production environments, prompt optimizing, testing and evaluation of AI systems, LLM application & agent development is desirable.
- Experience or publications in multi-modal is a plus, including image/video/audio understanding, text-to-image, text-to-video, etc.
Responsibilities
- Lead the investigation, exploration, and creation of next-generation, high-capacity LLM platforms and innovative products, while keeping close track with the latest AI technologies.
- Work closely with cross-functional teams to plan and implement projects harnessing LLMs for diverse purposes and vertical domains.
- Design and develop AI agents for a wide range of AI native applications, including multi-modality(image/video/audio).
- Continuously improve the agent's ability of understanding, reasoning, tool selection, taking actions at blazing fast speed.
- Develop and operate the big model service platform that offers businesses Model-as-a-Service solutions (MaaS).
- Research and develop advanced techniques for MaaS solutions including model post-training: fine-tuning (SFT), reinforcement learning (RL), reasoning, evaluation, test-time techniques, prompt optimization, multi-modality, and also LLM application/agent development.
Other
- Successful candidates must be able to commit to an onboarding date by end of year 2026.
- Please state your availability and graduation date clearly in your resume.
- Candidates can apply for a maximum of TWO positions and will be considered for jobs in the order you applied for.
- Maintain a deep passion for contributing to the success of large models is essential in this innovative and fast-paced team environment.
- Excellent problem-solving skills and a creative mindset to address complex AI challenges.
