Software Engineer - Machine Learning III

Samsung Research America seeks a Machine Learning Engineer to lead development of prompt injection and prompt safety models for agentic AI systems. You will design and deploy guardrail models using RLHF and DPO, and build hybrid on-device/cloud inference pipelines.

• Design and train prompt injection detection models and safety classifiers. Build hybrid deployment pipelines for on-device and cloud safety inference. Apply post-training techniques like RLHF and reward modeling. Curate adversarial training data. Build evaluation harnesses.

• M.S. or Ph.D. in CS/ML/EE or equivalent. 3+ years ML engineering experience. Strong Python and PyTorch skills. Hands-on experience with RLHF, DPO, and training classifier models for safety. Familiarity with prompt injection and adversarial robustness.

• Experience with agentic AI safety systems, red-teaming, on-device ML deployment, and publications in top venues.