Board Technology Adviser — Generative AI Governance, Banking

The mandate

The standing question is whether the bank can move from scattered generative-AI experiments to customer-facing deployment without accepting model, conduct and data risks the board cannot see. Directors want a voice that can distinguish durable capability from vendor theatre.

The adviser will reserve three days each month: one working session with the Chief Technology and Risk Officers, one portfolio review, and Technology Committee attendance. Material ad-hoc questions receive an initial view within one business day and a considered note within seventy-two hours.

The appointment runs for nine months and may be renewed once for three months by the committee chair after an eighth-month effectiveness review. The adviser holds no line authority, procurement vote or executive responsibility; management remains accountable for models, controls and delivery.

The retainer may sit alongside no more than two other substantial advisory commitments. A board role, investment or paid assignment with an Indian retail bank, large payments processor or a vendor being evaluated must be disclosed before interview and may constitute a conflict.

Why the board wants this voice

Management presents strong technical proposals, while risk discussions arrive after investment choices have hardened. The committee lacks a member who has taken probabilistic systems from lab to regulated production. An independent adviser can make uncertainty legible before capital and reputation are committed.

What you will own

• Test the assumptions behind the bank's twelve highest-value GenAI use cases and expose benefits that depend on unavailable data.
• Press the committee on risk appetite for hallucination, customer harm, intellectual-property leakage and human override.
• Shape a stage-gate that separates exploration, controlled pilot and production approval without becoming the approval authority.
• Challenge management's build, buy and partner logic for foundation models, orchestration layers and evaluation tooling.
• Guide the design of board metrics covering model drift, exception rates, unit economics and unresolved control breaches.
• Stress-test the proposed accountability map across technology, business, legal, compliance and model-risk teams.
• Advise the chair on questions to put to vendors before any multi-year platform commitment is authorised.

Candidate qualifications

• 22–28 years leading technology, data or model-risk decisions in a regulated financial institution.
• Direct governance experience with at least two machine-learning or GenAI deployments serving more than one million customers.
• Evidence of briefing a bank board or technology committee on material technology risk and investment trade-offs.
• Working command of model evaluation, retrieval architectures, privacy engineering and responsible-AI control patterns.
• Experience adjudicating a build-versus-buy decision involving global cloud or foundation-model providers.
• Independence from vendors likely to bid for the bank's GenAI platform or assurance work.

Non-negotiables

• Capacity for three scheduled days each month and one-business-day acknowledgement of urgent board queries.
• No current executive role, board seat or investment creating a conflict with the bank or shortlisted suppliers.
• Willingness to record dissent when the evidence does not support management's preferred technology bet.
• Bengaluru attendance for every scheduled committee meeting unless the chair agrees an exception in advance.

1. 49 words maximum. Which regulated GenAI deployment have you personally governed, and what evidence changed your recommendation to its board?
2. 49 words maximum. Which current boards, retainers, investments or vendor relationships would you disclose to this bank before appointment?
3. 49 words maximum. Can you protect three days monthly and respond to a material committee question within one business day for nine months?