Machine Learning Engineer - LLMs & Generative AI

Responsibilities

• Lead the development, training, and deployment of large language and multimodal foundation models tailored to clinical and biomedical domains.
• apply and refine state-of-the-art techniques such as supervised fine-tuning (SFT), reinforcement learning-based methods (e.g., RLHF, RLVR), parameter-efficient fine-tuning (PEFT), prompt tuning, and retrieval-augmented generation (RAG).
• Collaborate cross-functionally with researchers, clinicians, and engineers to design ML-driven solutions that improve healthcare delivery and outcomes.
• Build scalable infrastructure for distributed training of large models (TPU/GPU clusters, multi-node orchestration).
• Design and evaluate models for robustness, bias mitigation, factual consistency, and explainability in healthcare contexts.
• Stay current with the latest research in generative AI and contribute back to the community through publications and open-source initiatives.
• Minimum Qualifications
• 6+ years of experience in software engineering or machine learning (3+ years with a PhD).
• Experience designing and training LLMs or large-scale generative models (e.g., GPT, PaLM, LLaMA, Claude, Gemma).
• Deep expertise in NLP, sequence modeling, and transformer architectures.
• Proficient in Python and ML libraries such as PyTorch or TensorFlow; strong engineering skills in building scalable ML pipelines.
• Experience with RL-based fine-tuning (e.g., Reinforcement Learning from Human Feedback) and evaluation of generative systems.
• Proven ability to lead technical projects and collaborate across teams.
• Bachelor's degree in Computer Science, Engineering, or a related technical field
• Preferred Qualifications
• PhD or equivalent experience in Machine Learning, NLP, AI, or a related field.
• Experience in healthcare, biomedical informatics, or clinical data modeling.
• Familiarity with multi-modal foundation models (e.g., text-image, text-structured data) and cross-modal alignment techniques.
• Hands-on experience with vector databases, semantic search, and retrieval-based generation