AI Engineer

The AI (Artificial Intelligence) Engineer designs, builds and operates production-grade decision engines spanning machine learning (ML) models, deep learning (DL) models, large language models (LLMs) and decision models. The engineer trains, distils and serves proprietary LLMs on company-owned GPU infrastructure, extends a proprietary autonomous multi-agent platform and deploys systems on company premises, in air-gapped client environments and on Oracle Cloud Infrastructure.

The platform serves organisations whose operations depend on real-time, high-stakes decisions in regulated and often air-gapped environments. Models are measured on the operational outcomes they move, and domain rules and expert judgements are encoded into the knowledge graphs and evaluation sets that models and agents consume.

KEY RESPONSIBILITIES

• Own mission-critical ML systems from scoping through deployment and monitoring.
• Pre-train, post-train, distil and prune proprietary LLMs; evaluate on real cases before release.
• Integrate LLMs and decision models with strict safety and output validation.
• Build probabilistic forecasting models with calibrated uncertainty.
• Architect low-latency inference pipelines; build and administer GPU servers and inference engines.
• Construct knowledge graphs with dense and sparse retrieval and real-time pipelines.
• Monitor drift, bias, performance degradation and business KPIs beyond accuracy, with automated retraining.
• Establish guardrails, audit trails, explainability, access controls and data lineage.
• Communicate model risks to stakeholders; maintain model cards, playbooks and incident guides.

KNOWLEDGE AND SKILLS

General Requirements

• Bachelor's or Master's degree in Computer Science, Mathematics, Statistics, Data Science or a related field
• Professional working proficiency in English, and in Bahasa Indonesia or Mandarin, is required as this role engages directly with client stakeholders in Indonesia and mainland China for requirement workshops, technical discussions and solution presentations in their working language.
• At least four of: a production ML system; a proprietary LLM with measured gains; a knowledge graph of hundreds of thousands of entities; an inference service meeting a latency target on self-managed GPUs; a peer-reviewed publication or thesis; a top five percent international competition placement - Mandatory

Technical Skills

Candidates must demonstrate depth across all of the following areas, not a subset.

• ML and DL: gradient-boosted methods, probabilistic forecasting, deep learning for time series - Mandatory
• LLM pre-training, post-training, distillation and pruning - Mandatory
• Reinforcement learning agents; neuro-symbolic integration of LLMs with formal solvers - Mandatory
• Low-latency inference; model export and runtime optimisation; GPU server build and administration - Mandatory
• Containerisation and CI/CD on premises, air-gapped and on Oracle Cloud Infrastructure - Mandatory
• Real-time pipelines; knowledge graphs with dense and sparse retrieval; graph analytics at scale - Mandatory
• Drift and bias monitoring; guardrails, audit trails, explainability and access controls - Mandatory
• Unix-like administration across Linux and BSD variants for air-gapped deployment - Mandatory
• Rust, Go, or C/C++ with production code delivered - Mandatory

Domain and Delivery Skills

• Building libraries, frameworks and tooling onto proprietary training, inference and agent infrastructure - Mandatory
• Encoding operational rules and expert judgements as knowledge graphs, evaluation sets and agent tools - Mandatory
• Applied ML in a real-time, high-stakes operational domain, evidenced by delivered or published work - Mandatory