Head Of AI

Location: Singapore Role Overview TSC builds AI-native stakeholder and issue intelligence products for enterprise teams operating across public affairs, regulatory, reputational, strategic and external-risk environments. We are hiring a Head of AI to own the technical strategy and delivery of the intelligence layer behind Genie. This is a hands-on product and engineering leadership role, not an AI research-lab position. You will lead AI Engineers and Data Engineers, partner closely with Product and Platform Engineering, and turn fragmented, fast-changing public and proprietary data into reliable, explainable, secure and commercially valuable intelligence. Your mandate covers production AI systems, data foundations, evaluation and governance, unit economics, and team leadership. Success will be measured not by model novelty, but by customer value, reliability, delivery velocity and sustainable cost. The Mandate Own the technical strategy, systems, team and operating model required to turn TSC's data into reliable, differentiated and commercially valuable AI-powered intelligence. What You Will Own 1. AI Products and Intelligence Systems Define TSC's AI technical strategy across stakeholder mapping, classification, summarisation, entity resolution, risk detection, insight generation and workflow automation. Identify and frame high-value AI opportunities based on customer problems, decision workflows and TSC's differentiated data assets. Partner with Product to translate ambiguous customer needs into measurable product requirements, acceptance criteria and release sequencing. Define the appropriate technical approach for each use case, including deterministic software, machine learning, retrieval, LLMs, agents and human review. Own the end-to-end lifecycle of material AI capabilities, from data and evaluation design through deployment, monitoring, incident response and continuous improvement. 2. Evaluation, Reliability and Governance Build TSC's evaluation system for AI products, including golden datasets, regression tests, retrieval-quality checks, extraction and hallucination metrics, human-review thresholds and business KPIs. Define release-quality standards and ensure AI features are observable, testable, traceable and reversible. Establish governance for model and agent use, including prompt injection, data leakage, tool permissions, approval workflows, audit logs and customer-specific restrictions. Define policies for confidential customer data, workspace isolation, access controls, retention, deletion, data residency and secure use of external model providers. Create clear incident, rollback and escalation processes for AI-related quality, security and operational failures. 3. Data Foundations Own the data foundations behind TSC's intelligence products: ingestion, transformation, enrichment, entity resolution, labelling, retrieval, freshness, telemetry and feedback loops. Improve the quality and usability of fragmented public, proprietary and customer-specific data used by Genie. Ensure data pipelines and retrieval systems are reliable, secure, maintainable and fit for enterprise-facing use. 4. Architecture and AI Unit Economics Architect production AI systems using LLM APIs, specialised and open-source models, retrieval, deterministic orchestration and tool-using agents where appropriate. Make explicit decisions on when to use agents, when deterministic workflows are safer, and when human review is required. Own AI unit economics across model routing, caching, context-window discipline, token usage, latency, fallbacks and vendor trade-offs. Balance customer value, quality, speed, cost, security and maintainability in architecture and roadmap decisions. Maintain clear technical decision records and communicate trade-offs to executives and non-technical stakeholders. 5. Agentic Systems and AI-Enabled Engineering Establish repeatable agentic workflows in which AI systems can plan, act, validate, diagnose and revise within defined permissions, budgets and stopping conditions. Apply these patterns to customer-facing intelligence products and, where valuable, to internal software and data engineering workflows. Design agent workflows around evaluations, tests, logs, version-controlled artefacts, issue trackers, search and browser tools, sub-agents and human approval gates. Set permission boundaries, escalation paths, stopping criteria and cost limits so AI tools improve velocity without creating uncontrolled autonom…