LLM Engineer (Reinforcement Learning)

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI

Last checked: 1 hour agoStatus: Live

Job Summary

Design and improve LLM training pipelines that power generation models used in real services. Drive efficiency and enhance accuracy and stability by applying alignment approaches (e.g., PPO/GRPO/DPO) and building training structures that mitigate reward hacking and enable self-refinement. Develop foundational models that integrate with external knowledge and APIs, including learning how to select the right tools based on instruction type.

Location: South Korea

Workplace: Hybrid

Employment Type: Full time

Job Function: Education & Training

Seniority: Mid level

Key Responsibilities

•Improve the efficiency of LLM training processes using direct alignment methods such as PPO/GRPO/DPO.
•Increase the accuracy and stability of generated outputs through improved training structures.
•Design training approaches to prevent reward hacking and enable self-refinement.
•Develop base models that can integrate external knowledge and APIs.
•Train LLMs to select the appropriate external tools based on instruction types.

Key Requirements

•3+ years of experience in deep learning or NLP (master’s new grads also eligible).
•Strong programming skills in Python and PyTorch.
•Experience designing, training, evaluating, and optimizing models using PyTorch.
•Experience training LLMs with GPUs, including troubleshooting.
•Experience using distributed training frameworks such as Slurm, DDP, and Horovod.

Experience:3+ yearsDeep learningNLPLLM

Skills:Collaboration

Tech Stack:PythonPyTorchSlurmDDPHorovodDockerKubernetesGPUPPOGRPODPODirect Alignment AlgorithmParameter Efficient Fine-TuningSupervised Fine-Tuning

Company Brief

42dot

Website

Loading job details...

Preparing the role view and application actions.

42dot

1 week ago

LLM Engineer (Reinforcement Learning)

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI

Last checked: 1 hour agoStatus: Live

Job Summary

Location: South Korea

Workplace: Hybrid

Employment Type: Full time

Job Function: Education & Training

Seniority: Mid level

Key Responsibilities

•Improve the efficiency of LLM training processes using direct alignment methods such as PPO/GRPO/DPO.
•Increase the accuracy and stability of generated outputs through improved training structures.
•Design training approaches to prevent reward hacking and enable self-refinement.
•Develop base models that can integrate external knowledge and APIs.
•Train LLMs to select the appropriate external tools based on instruction types.

Key Requirements

•3+ years of experience in deep learning or NLP (master’s new grads also eligible).
•Strong programming skills in Python and PyTorch.
•Experience designing, training, evaluating, and optimizing models using PyTorch.
•Experience training LLMs with GPUs, including troubleshooting.
•Experience using distributed training frameworks such as Slurm, DDP, and Horovod.

Experience:3+ yearsDeep learningNLPLLM

Skills:Collaboration

Tech Stack:PythonPyTorchSlurmDDPHorovodDockerKubernetesGPUPPOGRPODPODirect Alignment AlgorithmParameter Efficient Fine-TuningSupervised Fine-Tuning

Company Brief

42dot

Website