Job Description

Labelbox is seeking an Applied Research Intern to design, build, and productionize evaluation and post-training systems for frontier LLMs and multimodal models. The intern will focus on creating high-quality evaluations, datasets, and prototyping training loops to improve AI performance, with the goal of integrating research into Labelbox products and collaborating with customers.

Responsibilities

• Build and own evaluation and benchmark suites
• Create post-training datasets
• Prototype RLHF/RLAIF/RLVR/RM/DPO-style training loops
• Ship improvements into Labelbox workflows
• Engage with customer research teams

Requirements

• Ph.D. or Master’s degree in Computer Science, Machine Learning, AI, or related field
• Understanding of autoregressive and diffusion multimodal models
• Passion and experience for LLM evaluation
• Expertise in training data quality
• Ability to bridge research and application
• Track record of publishing in top-tier AI/ML conferences
• Proficiency in Python and deep learning frameworks

Benefits

• No benefits