AI/ML Core Modeling Engineer, New PHD Grad

gNucleus AI is an end-to-end Generative AI platform focused on engineering design and simulation. The AI/ML Core Modeling Engineer will own the post-training stack for adapting open-source foundation models to enterprise domain-specific problems, including dataset curation, supervised fine-tuning, reinforcement learning, inference optimization, and evaluation. The role also involves collaborating with cross-functional teams to improve ML system performance and efficiency.

Responsibilities:

• Filter, structure, and synthetically generate high-quality datasets (SFT/RL) tailored to specific customer domain data and workflows
• Lead Supervised Fine-Tuning (SFT) and Reinforcement Learning experiments (e.g., DPO, GRPO, PPO, KTO) to adapt base models to customer tasks, reasoning needs, and safety constraints
• Optimize and serve post-trained models for low-latency, high-throughput deployment (using engines like vLLM, SGLang, or TensorRT-LLM) to meet customer production SLAs
• Build bespoke evaluation pipelines, custom benchmarks, and LLM-as-a-judge frameworks to measure domain accuracy, alignment, and alignment progress for customer review
• Collaborate with cross-functional teams to design and optimize ML systems, leveraging expertise to drive performance improvements and efficiency gains