Amay Kalal

• Shaped the AI governance strategy for 3 Tier-1 banks ahead of EU AI Act enforcement, defining risk classification
hierarchies, conformity assessment processes, and Fundamental Rights Impact Assessment (FRIA) methodologies
that were adopted as the enterprise standard across each institutions AI programme.
• Established the Model Risk Management architecture for 5 production LLM systems, covering model cards,
explainability requirements, human oversight controls, and audit-ready governance artefacts, achieving a 100%
pass rate across all regulatory examination reviews and positioning clients for sustained regulatory confidence.
• Reduced model risk incidents by 30% before production deployment by introducing LLM-as-a-Judge evalua
tion frameworks with Explainable AI (XAI) methodology, tracking hallucination rates, bias scores, and ethical
alignment, and shifting client culture from reactive incident response to proactive AI risk governance.
• Advised C-suite and board-level risk committees on responsible AI adoption, translating complex LLM system
risk, agentic AI behaviour, and regulatory change across EU AI Act, NIST AI RMF, and ISO 42001 into governance
roadmaps and remediation plans that secured executive alignment and investment approval.
• Built and led a team of 6 AI governance and engineering professionals across the full programme lifecycle, driving
scope governance, stakeholder management, and executive reporting for Tier-1 financial clients with consistent
on-time, on-budget delivery.

• Defined the AI quality and compliance governance standard adopted across 3 programme streams serving US
and European banking clients, with 500+ validated controls embedded into delivery workflows, achieving 30%
improvement in delivery predictability and 95% compliance pass rates across all CI/CD pipelines.
• Led compliance and security assessments across 4 open banking and PSD2 payment systems for global financial
clients, identifying and remediating authentication gaps, data handling risks, and regulatory exposure before
production release, materially reducing client regulatory risk.
• Delivered responsible AI operational standards for banking infrastructure serving 10M+ customers, embedding
post-market monitoring, observability, and human oversight mechanisms that sustained 99.9% uptime and met
emerging AI operational governance requirements.