Model Inference Architect — High-Performance AI Systems

YTL AI Labs; Kuala Lumpur; Vor Ort; MYR 180.000 - 320.000; Vollzeit

Zusammenfassung

YTL AI Labs is seeking a Model Performance Engineer in Kuala Lumpur to lead the design and optimization of large-scale AI inference systems. This role sits at the crossroads of machine learning research and distributed systems engineering, driving how frontier models are deployed, scaled, and experienced in production.

You will shape the inference architecture, pushing latency, throughput, and cost efficiency, translating cutting-edge research into robust, real-world systems.

Qualifikationen

• 3–5+ years in ML systems, backend engineering or HPC.
• Strong with deep learning frameworks (PyTorch, JAX, TF).
• Proficient in Python and/or C++.
• Knowledge of Transformer LLMS and GPU parallelism.
• Experience with production-grade inference systems.

Aufgaben

• Design and implement high-performance inference systems for LLMs, multimodal, and speech models.
• Drive improvements in throughput, latency, and cost efficiency.
• Architect and optimize distributed inference across GPU clusters.
• Own techniques like KV-cache optimization and memory management.
• Evaluate serving frameworks (vLLM, TensorRT-LLM, Triton, custom runtimes).
• Partner with research to productionize new models and architectures.
• Build benchmarking and evaluation pipelines for inference performance.
• Diagnose bottlenecks in networking and distributed systems.
• Mentor engineers and contribute to technical direction and best practices.

Kenntnisse

ML systems

Backend engineering

High-performance computing

Deep learning

Python

C++

Transformer models

GPU architecture

Distributed systems

CUDA memory mgmt

Tools

vLLM

TensorRT-LLM

Triton

Custom runtimes

Jobbeschreibung

YTL AI Labs is seeking a Model Performance Engineer in Kuala Lumpur to lead the design and optimization of large-scale AI inference systems. This role sits at the crossroads of machine learning research and distributed systems engineering, driving how frontier models are deployed, scaled, and experienced in production.

You will shape the inference architecture, pushing latency, throughput, and cost efficiency, translating cutting-edge research into robust, real-world systems.