LLM Research & Validation Specialist

About the role

Lead frontier research, quantitative evaluation and independent validation of large language models, multimodal models, retrieval‑augmented generation systems and agentic AI for ADIB. Translate scientific methods into reproducible tests and evidence for model governance.

Key responsibilities

• Design and run reproducible evaluation harnesses, golden datasets, adversarial suites and statistically defensible acceptance criteria for LLMs and related AI systems.
• Assess model architecture, tokenisation, embeddings, attention, decoding, fine‑tuning, alignment, quantisation and inference configurations.
• Perform deep testing of prompt injection, data leakage, tool‑use safety, agency risks and human‑oversight mechanisms.
• Build and maintain Python‑based validation tooling, automated test pipelines and experiment‑tracking repositories.
• Prepare validation reports, research notes and senior‑management briefings that clearly separate evidence, judgement and residual uncertainty.

Required profile

• Master’s degree in a quantitative discipline (Physics, Mathematics, etc.); PhD or research‑intensive master strongly preferred.
• 1‑3 years of experience in AI research, machine learning, quantitative modelling or model validation.
• Strong understanding of probability, statistics, linear algebra, optimisation and uncertainty quantification.
• Ability to read and reproduce research papers, design controlled experiments and communicate findings to non‑technical stakeholders.

Required skills

• Advanced Python programming.
• Experience with PyTorch and Hugging Face libraries.
• SQL and Git version control.
• Familiarity with cloud AI platforms and evaluation frameworks.