大模型算法专家 - 豆包输入法
Core
Develop and optimize large language models for a next-gen AI input method, focusing on speech recognition, intelligent suggestion, and keyboard input accuracy.
Role type
Senior IC large model algorithm engineer (input methods)
Builds
AI input system features for mobile and desktop platforms
Domain
Consumer AI / NLP / Input Methods
Deliverable
production ML models
Required skills
SFT, DPO, RL, KD, LoRA, QLoRA, Adapter Tuning, PyTorch, Hugging Face Transformers, MLOps, model evaluation, A/B testing
Preferred skills
NeurIPS/ICML/ICLR/ACL/EMNLP publications, ACM/ICPC/NOI/IOI/Top Coder/Kaggle awards, experience in spelling/autocomplete/handwriting/speech/multimodal domains
Technologies
PyTorch, Hugging Face Transformers
Responsibilities
Research and develop post-training algorithms (SFT, DPO, RL, KD); design model architecture and data pipelines; collaborate with product/ops to implement model capabilities; explore and apply frontier LLM technologies.
Seniority
Senior, hands-on IC