Vice President - Senior Core AI Engineer

SGX Group; Singapore; On-site; SGD 180,000 - 280,000; Full time

Job summary

SGX Group is seeking a Core AI Engineer in Singapore to build the shared platform for SGX AI use cases. You will work on the enterprise knowledge graph, governed retrieval and reusable AI components used by business units.

The role emphasizes encryption, access control, traceability and high-assurance engineering within a regulated, confidentiality-critical environment.

Qualifications

• Proven software engineering experience incl. system design, testing, debugging and secure coding.
• Hands-on coding in Python and TypeScript; experience with Rust is a plus.
• Experience building applications with LLMs: RAG and KAG pipelines, embeddings and semantic retrieval.
• Experience modelling connected data: knowledge graphs, ontology design and graph queries.
• Strong SQL and PostgreSQL; vector search experience with pgvector.
• Experience designing and maintaining APIs and shared services; versioning and backward compatibility.
• Applied cryptography and secure data handling; AES-256-GCM, envelope encryption, KEK/DEK, cloud KMS or HSM.
• Ability to threat-model data paths carrying secret data; understanding TLS limitations.

Responsibilities

• Build and operate the core knowledge graph and knowledge-management layer; ontology, metadata, and governed ingestion.
• Develop shared platform features: APIs, tool registry, MCP interfaces, reusable AI components.
• Ensure end-to-end retrieval with permission controls, labeling, and filtering in vector search and graph traversal.
• Institute traceability: citations, transformation lineage, content hashing and tamper evidence, audit trails.
• Manage encryption and key management for confidential data with per-record keys and rotation.
• Define evaluation harness: golden task sets, scoring rubrics, reproducible quality and cost claims.
• Design observability, monitoring and cost governance for AI services; optimize business value.
• Define APIs and architecture; present designs for Architecture Forum, ARB and InfoSec readiness.
• Translate recurring patterns into platform capabilities; support engineers building on them.
• Lead pre-deployment reviews for medium/high-risk use cases; verify vendor work in SGX environments.

Skills

Python

TypeScript

LLMs

Knowledge graphs

Graph databases

SQL

PostgreSQL

API design

Cryptography

Threat modeling

Cloud security

Architecture decisions

Private cloud

Tools

pgvector

Cypher

Rust

KMS

Job description

SGX Group operates Singapore's securities and derivatives exchange. The Core AI Engineer is a senior individual contributor role within the AI & Automation Engineering team, building the shared platform that every SGX AI use case runs on: the enterprise knowledge graph and knowledge-management layer, the governed retrieval and access services above it, and the reusable AI components, tools and APIs that engineers embedded in the business units consume. The platform handles confidential and market-sensitive material, so permission-aware retrieval, encryption and key management, traceable answers and measured quality are product requirements rather than afterthoughts. This is a Singapore-based role.

Job description:

Job Summary

SGX Group operates Singapore's securities and derivatives exchange. The Core AI Engineer is a senior individual contributor role within the AI & Automation Engineering team, building the shared platform that every SGX AI use case runs on: the enterprise knowledge graph and knowledge-management layer, the governed retrieval and access services above it, and the reusable AI components, tools and APIs that engineers embedded in the business units consume. The platform handles confidential and market-sensitive material, so permission-aware retrieval, encryption and key management, traceable answers and measured quality are product requirements rather than afterthoughts. This is a Singapore-based role.

Job Responsibilities

• Build and operate the core knowledge graph and knowledge-management layer: ontology and metadata model, governed ingestion from enterprise sources, graph projection, and hybrid graph and vector retrieval
• Build the shared platform features other teams consume: retrieval and access-layer APIs, a tool registry and MCP interfaces, and reusable AI components that stay framework-agnostic so any consuming team can build on them
• Make retrieval permission-aware end to end: sensitivity labels inherited at ingestion, closed-world enforcement where an unlabelled record is denied rather than defaulted, filtering applied inside vector search and graph traversal, and live validation on the sources actually cited
• Engineer traceability and integrity into the platform: citations, transformation lineage from source document to chunk, content hashing and tamper evidence, and an append-only audit of every access decision and write
• Own encryption and key management for confidential and market-sensitive data: envelope encryption with per-record data keys wrapped by a KMS-held key, authenticated encryption at rest and on the wire, client-side decryption where the service itself must not see plaintext, and key rotation, re-wrapping and revocation
• Own the evaluation harness: define golden task sets and scoring rubrics, benchmark approaches head to head, and make quality, cost and latency claims reproducible rather than anecdotal
• Design and implement observability, monitoring and token and cost governance for AI services, and propose fit-for-purpose solutions that optimize business value without overextending the AI budget
• Define APIs, integration patterns, cohesive architecture, and engineering standards; take designs through Architecture Forum, ARB and InfoSec Release Readiness, and own the resulting technical and governance artefacts
• Turn recurring patterns from business-unit use cases into platform capabilities, and support the engineers who build on them
• Lead pre-deployment reviews for medium and high-risk use cases, and technically review work delivered by vendors and partners, reproducing their results in an SGX environment before accepting them

Job Requirements

• Proven software engineering experience, including system design, code review, testing practices, debugging in production environments, performance tuning, secure coding, and building reliable, maintainable services
• Strong hands-on coding in Python and TypeScript, and comfort working polyglot, including reading and writing a systems language (parts of the platform are Rust)
• Hands-on experience building applications with LLMs: RAG and KAG pipelines, document processing, embeddings and semantic retrieval, and agent and tool integration
• Experience modelling connected data: knowledge graphs, ontology and taxonomy design, entity and relationship extraction, and graph query languages such as Cypher, or deep relational data modelling with a clear appetite to work in graph
• Strong SQL and PostgreSQL, including schema design and query performance, and experience with vector search such as pgvector
• Experience designing and maintaining APIs and shared services that other engineering teams depend on, including versioning and backward compatibility
• Strong applied cryptography: correct use of authenticated encryption such as AES-256-GCM including nonce discipline, envelope encryption with a KEK and DEK split, key management through a cloud KMS or HSM, key rotation and re-wrapping, and careful handling of key material in memory
• Able to threat-model a path that carries secret data and write down what is and is not covered, including where TLS alone is insufficient — such as a TLS-terminating corporate proxy — and what metadata leakage means when the content is market-sensitive
• Experience applying data governance and security controls in regulated environments: data classification, least-privilege access, secrets management with owned rotation, keyless federated credentials, and audit logging
• Proven ability to make and document architectural decisions with minimal oversight, and to defend them in an architecture or security review
• Comfortable delivering inside a locked-down cloud environment: private-only networking, egress allow-lists, keyless federated credentials, no local container builds, and schema and deployment changes applied under review

Preferred Skills:

• Experience in financial services or regulated industries
• Familiarity with Google Cloud Platform and Gemini APIs
• Experience with agent orchestration frameworks
• Background in MLOps or AI/ML platform engineering
• Experience with graph databases or vision-language models
• Experience in financial services, capital markets, or another regulated industry
• Familiarity with Google Cloud Platform and Vertex AI / Gemini APIs, including Cloud Run, Pub/Sub and Cloud SQL
• Neo4j or another graph database in production, and graph RAG or hybrid graph and vector retrieval
• Hands-on with cryptographic libraries and protocols: AEAD constructions, ECDH key agreement such as X25519, HKDF key derivation, Cloud KMS or HSM-backed keys, and Web Crypto for browser-side decryption
• Having taken an encryption or key-management design through a security review in a regulated firm
• Agent orchestration frameworks and Model Context Protocol (MCP), such as Google ADK, Claude Agent SDK or LangGraph
• Enterprise connector and ingestion work, such as Microsoft Graph / SharePoint, Confluence, and event-driven pipelines
• LLM evaluation practice: LLM-as-judge, deterministic scoring, and regression suites for prompts and models
• Background in MLOps or AI/ML platform engineering, and effective use of AI coding agents to raise delivery throughput
• Vision-language models or document AI for complex document understanding