Graison Thomas

Background

About Graison Thomas

I am a seasoned Model Risk and AI/ML specialist with 16+ years of experience across leading financial institutions, including American Express, Credit Suisse, Citibank (Consultant), Bank of America (Consultant), and Rabobank (Consultant). My career spans model development, validation, and internal audit, providing an end-to-end understanding of model risk management, AI-driven innovation, and regulatory compliance (SR 11-7, CCAR, CECL, Basel). My technical expertise includes traditional statistical models, advanced machine learning techniques, and large language models (LLMs), with a strong focus on deploying these technologies responsibly in highly regulated environments. I’m particularly passionate about the evolving risks and opportunities posed by Generative AI and foundation models in financial services: & – Extensive experience in assessing model performance, robustness, and interpretability to ensure fitness-for-use and alignment with regulatory expectations./ & – Hands-on expertise with NLP models such as GPT, LLaMA, and BERT-based architectures; boosting algorithms like XGBoost and LightGBM; and deep learning frameworks including Transformers, CNNs, and RNNs/LSTMs. – Proficient in knowledge distillation, transfer learning, finetuning, hyperparameter tuning, model explainability, and transformer-based architectures. – Proven track record in developing and validating models for Credit Risk, Fraud Detection, Marketing, PPNR, Capital Planning, and Market Risk. & – Experienced in leading and mentoring cross-functional teams and fostering a culture of continuous learning and excellence. & : Master’s in AI & ML | MBA | B.E. in Electronics FRM | CQF | CFA Level II As financial institutions adopt increasingly sophisticated ML models and LLMs, I focus on ensuring innovation is paired with strong governance, regulatory alignment, and ethical deployment. My goal is to help build resilient, risk-aligned, and scalable AI solutions that enhance decision-making while mitigating risk.