Senior Director, AI-Ready Data & Harness Engineering

An empowering career at Singtel begins with a Hello. Our purpose, to Empower Every Generation, connects people to the possibilities they need to excel. Every "hello" at Singtel opens doors to new initiatives, growth, and BIG possibilities that takes your career to new heights. So, when you say hello to us, you are really empowered to say…“Hello BIG Possibilities”.

A deep-tech, hands-on AI/data engineering leader to head Singtel's AI-Ready Data & Harness Engineering pillar in AIDA 2.0. Reporting directly to the Chief AIDA Officer, this role owns the design, build, run and continuous improvement of AI-ready reusable data products, knowledge/context assets, agent memory capabilities, retrieval/grounding harnesses and AI data-readiness governance at Singtel scale.

At Singtel scale, this pillar is the foundational AI data layer that creates the multiplier effect in AI returns: common data products, semantic context, retrieval and memory patterns should be built once and deployed multiple times across agents, models, journeys and BUs.

Make an impact by

AI-Ready Data & Harness Strategy and CXO-1 Deep-Tech Accountability

• Own the AI-Ready Data & Harness Engineering strategy, technical roadmap and capability architecture for Singtel SG, aligned to the AIDA 2.0 stack and Chief AIDA Officer agenda.
• Operate as a CXO-1 technology leader: shape, challenge and co-own senior decisions on AI-ready data investments, data architecture, knowledge/context engineering, governance, sequencing, technical trade-offs and outcomes.
• Translate enterprise AI ambition into reusable data products, knowledge assets, context/memory capabilities, retrieval harnesses, evaluation assets, data standards and delivery playbooks.
• Treat the AI-ready data and harness layer as the foundational multiplier for AIDA returns, explicitly designing for build-once, deploy-many reuse across models, agents, channels and BUs.
• Set the bar for AI-ready data product design, data contracts, semantic consistency, trusted context, privacy/security-by-design, operational reliability and measurable business value.
• Ensure Singtel's AI investments are constrained by value and quality - not by data fragmentation, unclear ownership, weak governance, stale context or brittle retrieval patterns.

Build and Own the AI-Ready Data Engineering Powerhouse

• Lead Tech FTE organization across AI-ready reusable data products, Knowledge Engineering, Context Engineering, Agent Memory Management and AI Data Readiness Governance.
• Build and coach high-calibre data product engineers, data engineers, data architects, knowledge engineers, ontology/semantic architects, context/RAG engineers, memory engineers, governance specialists and AI data operations talent.
• Create a culture of product ownership, reusable design, engineering depth, data quality, governance-by-design, production reliability, cost discipline and business-value accountability.
• Use partners and specialist vendors selectively while retaining ownership of critical data IP, architecture, standards, semantic definitions, governance, retrieval quality and delivery accountability.
• Raise the expertise level of existing data-readiness capabilities and rationalize fragmented or duplicated work into stronger, accountable sub-functions.

Reusable AI-Ready Data Products for All BUs

• Own the portfolio of reusable AI-ready data products across Consumer, Enterprise, Network & IT and corporate functions, aligned to the AIDA 2.0 org expectation for all BUs.
• Define productization standards for data domains, data contracts, APIs, metadata, discoverability, ownership, quality thresholds, lineage, access controls, SLAs/SLOs, lifecycle and retirement.
• Convert fragmented datasets, documents, events and operational signals into certified, reusable assets that AI / ML, GenAI and Agentic AI teams can consume through stable interfaces.
• Prioritize high-value data products that accelerate multiple AI use cases rather than one-off pipelines for isolated pilots.
• Design priority data products for build-once, deploy-many scale, maximizing reuse across BUs and AI journeys instead of creating bespoke pipelines or duplicated domain assets.
• Track adoption, reuse, freshness, quality, cost-to-serve, defect rates, time-to-integrate and business value contribution for each major data product.

Knowledge Engineering, Ontology, Knowledge Graphs and Semantic Layer

• Own Knowledge Engineering capabilities (with federated operational structure) including ontology, taxonomy, entity resolution, master/reference data alignment, business glossary, knowledge graphs and semantic layers.
• Build enterprise intelligence assets that link customer, product, service, network, device, order, ticket, interaction, finance, campaign, process and policy knowledge where relevant for AI value.
• Ensure semantic consistency across BUs so models, agents, dashboards and decision services use common definitions, relationships and trusted business meaning.
• Establish trusted single-ledger semantics for customer, product, employee, network and operational entities, especially where UDP and the Network on-prem HP lake do not yet provide an enterprise semantic layer.
• Partner with AI Capabilities & Services, Central AI Kitchen and BU teams to expose knowledge assets through APIs, graph queries, semantic services and reusable context packages.
• Design knowledge assets for explainability, auditability, policy enforcement, search/retrieval quality and responsible AI use.

Context Engineering, Retrieval Harnesses and Agent Memory Management

• Own Context Engineering (with federated operational structure) for AI/GenAI/Agentic AI across chunking, embeddings, vector stores, graph retrieval, hybrid search, ranking, prompt/context packaging, caching, tool data interfaces and context-window economics.
• Build reusable RAG and retrieval harnesses with evaluation datasets, gold-standard answers, grounding checks, retrieval-quality metrics, regression tests, traceability and improvement loops.
• Own Agent Memory Management patterns for short-term and long-term memory, user/session/entity/process memory, memory write/read policies, retention, privacy, explainability and safety controls.
• Ensure agents consume consistent governed context and memory so hundreds of agents do not form divergent customer/product understanding; optimize context availability, freshness and retrieval paths for low-latency customer experience.
• Diagnose context and retrieval failures with data scientists, agent engineers and business SMEs, including stale sources, missing entities, poor chunking, weak metadata, bad rankings, hallucination-inducing gaps and latency/cost trade-offs.
• Partner with AI & Agent Ops for production telemetry, feedback loops, incident response, rollback, re-indexing, refresh cycles and post-launch improvement of context and memory capabilities.

AI Data Readiness Governance, Trust and Policy-by-Design

• Own AI Data Readiness Governance for AIDA, covering quality, discoverability, lineage, provenance, policy, ownership, access, classification, privacy, consent, retention and auditability.
• Define certification gates for experimentation, pilot, production launch and scale-up so AI use cases consume trusted, policy-compliant and fit-for-purpose data/context assets.
• Partner with DPM/data owners, IT/CIO, Cyber/CISO, legal/regulatory and business teams to make data governance an accelerator for AI delivery rather than a late-stage blocker.
• Implement governance-by-design in data products, knowledge assets, retrieval stores, agent memory, logs, feedback data and model/agent evaluation datasets.
• Maintain clear policies for sensitive data, customer data, operational data, third-party data, generated data, human feedback, agent traces and derived intelligence assets.

Production Data and Harness Delivery Without Data Debt

• Develop and operationalize high-priority AI-ready data products, knowledge/context assets and harness capabilities with AIDA Business Partners and business teams.
• Convert ambiguous business problems into data product designs, source-system/integration requirements, semantic models, retrieval architectures, governance plans, adoption paths and measurable outcomes.
• Embed AI-ready data and harness capabilities into Central AI Kitchen, AI Services & Capabilities, AI & Agent Ops, business workflows, enterprise systems and decision processes.
• Ensure data and context services are secure, scalable, observable, testable, resilient, cost-efficient, maintainable and supportable by Day 2 operations.
• Make deliberate trade-offs that accelerate value while avoiding brittle one-off pipelines, hidden data debt, unmanaged dependencies, duplicate semantic layers, ungoverned shadow datasets and fragile run operations.

Ecosystem Orchestration, Critical Path Acceleration and Engineering Culture

• Coordinate with AIDA, IT, Cyber, data owners, vendors, hyperscalers and industry partners to co-solve emerging semantic/context/harness patterns while avoiding premature lock-in and preserving speed-to-value.
• Crash critical paths by surfacing data/source-system dependencies early, clarifying ownership, forcing data/architecture/governance decisions, removing blockers and escalating trade-offs at the right level.
• Build operating rhythms for data product prioritization, source-system readiness, governance review, context quality, semantic alignment, cyber review, platform integration, release readiness and post-launch improvement.
• Translate technical data and harness trade-offs into clear executive choices while retaining credibility with expert data engineers, AI engineers, architects, cyber teams and SMEs.

Skills for Success

A senior deep-tech AI-ready data and harness engineering leader with:

• 20+ years of hands-on experience across enterprise data engineering, data products, analytics platforms, AI/ML data foundations, MLOps/LLMOps data integration, knowledge/context engineering, retrieval systems and AI data governance.
• Experience building and owning enterprise-scale AI-ready data, knowledge/context, data platform or AI-enablement organizations at Singtel scale or higher, ideally 30-50+ data engineers, data product engineers, knowledge engineers, AI platform engineers and governance specialists.
• Direct involvement in technical design and delivery: data architecture, batch/stream pipelines, APIs, data contracts, metadata, quality, lineage, semantic layers, ontologies, knowledge graphs, feature/embedding/vector stores, RAG/retrieval harnesses and agent memory patterns.
• Strong track record creating AI-ready reusable data products and data governance capabilities that are adopted by multiple BUs and production AI/agent programs, not just dashboard/reporting datasets.
• Hands-on performance improvement with data scientists, AI engineers, data engineers and business SMEs, including root-cause analysis of data quality, freshness, semantic ambiguity, retrieval misses, missing context, memory errors, latency, cost and workflow failures.
• Track record delivering large, multi-stakeholder AI / data / digital programs from strategy through production launch, adoption, operations and measurable value realization.
• Ownership of ROI, investment cases, data product economics, reuse targets, adoption metrics, productivity outcomes, EBIT contribution and technical/data debt management.
• Experience coordinating across AIDA, IT, Cyber, DPM/data governance, data owners, platform teams, product, business, vendors and partners in regulated enterprise ecosystems.
• Proven ability to identify new opportunities, prioritize fewer/bigger bets, crash critical paths and unblock delivery without compromising architecture, governance, security, reliability or maintainability.
• Enterprise data engineering and data product architecture across batch/stream pipelines, APIs, data contracts, data quality, metadata, lineage, ownership, observability, lifecycle and SLA/SLO management.
• AI-ready reusable data products across customer, product, service, network, operations, sales, finance and enterprise domains, with strong product management, adoption, reuse and cost-to-serve discipline.
• Knowledge Engineering across ontologies, taxonomies, entity resolution, business glossaries, semantic layers, knowledge graphs, graph querying and enterprise intelligence asset design.
• Context Engineering and RAG/retrieval across chunking, embeddings, vector stores, hybrid search, graph retrieval, ranking, prompt/context packaging, grounding, evaluation datasets and regression testing.
• Agent Memory Management across user/session/entity/process memory, read/write policies, retention, privacy, explainability, security, observability and production improvement loops.
• AI Data Governance & Trust across quality, discoverability, policy, privacy, consent, access control, classification, retention, provenance, auditability, responsible AI and model/data risk.
• Production integration with AIDA stack: AI capabilities, Central AI Kitchen Platform & Ops, AI & Agent Ops, AI Value Realization Office, IT architecture, Cyber, DPM/data owners and BU systems/workflows.
• Executive gravitas to operate as a Chief AIDA Officer direct report and influence CXO-level business, technology, cyber, IT, DPM/data and governance leaders.
• Roll-up-sleeves technical leadership style; comfortable moving from ExCo-level trade-offs into detailed design reviews with data engineers, data product owners, knowledge engineers, AI engineers, platform engineers and cyber teams.
• Decisive prioritization and trade-off capability; able to sharpen focus, crash critical paths and make clear calls under ambiguity.
• Strong ecosystem orchestration across AIDA, IT, Cyber, DPM/data owners, platform, governance, product, business teams, vendors and partners.
• Commercial and ROI discipline; able to connect data/harness choices to adoption, productivity, revenue enablement, EBIT, compute/model economics and cost-to-serve outcomes.
• Talent builder and culture shaper who can attract scarce data/AI engineering talent, raise engineering standards and create a high-accountability AI-ready data engineering organization.
• Ability to balance delivery speed with long-term engineering integrity, avoiding fragile pilots, duplicate data products, unmanaged dependencies, ungoverned datasets and data debt.
• Ability to communicate complex data, governance, retrieval and AI harness trade-offs in business language while maintaining technical credibility with expert teams.