Vaibhav Singh
I design and deploy end-to-end Generative AI, LLM, and RAG pipelines. Specialized in productionizing scalable AI platforms and robust evaluation systems. Co-authored 2 AI safety publications (CVPR, EMNLP).
Impact & Leadership
Led a team of 4-5 engineers to productionize scalable AI platforms. Built automated evaluation pipelines and ran rapid experiments to drive product maturity.
System Specs
Generative AI Core
Model Engineering
MLOps & Infrastructure
Research Publications
Immune: Improving Safety Against Jailbreaks
Proposed an inference-time defense framework leveraging safe reward models to defend against jailbreak attacks in Multi-modal LLMs.
RELIC: Reward Model Generalization
A novel framework for enhancing reward model generalization in low-resource settings using few-shot examples.