AI Engineer (Model) at Toss

Department: 토스코어

Location: United Kingdom
Level: not_specified
Posted: Jun 9, 2026

Description

Toss is hiring an AI Engineer (Model) to build and operate in-house LLM and VLM models for real-world business problems. The role covers data creation, training pipelines, post-training techniques, evaluation benchmarks, and continual learning, with a focus on performance, latency, and cost. Candidates should have experience running the full model cycle, designing experiments, optimizing training pipelines, and applying techniques such as SFT, DPO, RL, LoRA, and quantization.

Required Skills

Data creation
Model training
Evaluation systems
SFT
DPO
RL
LoRA
Quantization
Distributed training
Synthetic data
LLM evaluations

Additional Information

Required Previous Experiences: Experience training LLM/VLM models and running the full data generation, training, and evaluation cycle
Experience designing experiments and quantitatively analyzing model results
Experience optimizing training pipelines for rapid modeling cycles
Experience with post-training techniques including SFT, DPO, RL, LoRA, and quantization
Preferred Qualifications: Experience with distributed training using DeepSpeed, FSDP, or Megatron
Experience designing synthetic data pipelines and analyzing data quality impact
Experience designing and automating LLM evaluations and benchmarks
Experience training and deploying domain-adapted models
Experience resolving issues in open-source LLM projects or contributing to the community
Translation Company Name: Toss