Travel LLM Engineer – Scalable ML Systems, Flexible Work

Enfint; Amsterdam; Hybrid; EUR 95,000 - 140,000; Full time

Benefits offered by this job

Parental leave 22 weeks

Bike reimbursement

Headspace membership

Employee Assistance Program

Discounts on Booking.com products

Job summary

Booking.com is seeking an ML Systems Engineer to design and maintain scalable training and inference infrastructure for large language models and data-intensive systems.

You will collaborate with ML scientists to productionize research, optimize serving, and build robust pipelines with monitoring and observability across the stack at our Amsterdam HQ, with flexible hybrid work options.

Qualifications

• Strong software engineering experience with ML systems at scale.
• Hands-on with LLM training infra and distributed training.
• Proficiency in Python and cloud infrastructure.
• Experience with Kubernetes, Docker, AWS/GCP, and ML tooling.

Responsibilities

• Design and maintain scalable ML training and fine-tuning infra.
• Improve model serving latency, throughput, and cost efficiency.
• Develop ML data pipelines, evaluation, and deployment.
• Implement optimization techniques for production models.
• Build monitoring and observability for ML systems.
• Collaborate with ML scientists to productionize research prototypes.
• Contribute to reusable ML frameworks and libraries.
• Ensure code quality through testing and reviews.

Skills

Python

Kubernetes

Docker

PyTorch

Distributed training

Cloud infrastructure

English communication

CI/CD for ML

Spark/Ray

Education

BSc in CS/Software Engineering

MSc in CS/Quant field

Tools

Kubernetes

Docker

AWS

GCP

PyTorch

DeepSpeed

Triton Inference Server

ONNX

TensorRT

vLLM

Spark

Ray

Job description

Booking.com is seeking an ML Systems Engineer to design and maintain scalable training and inference infrastructure for large language models and data-intensive systems.

You will collaborate with ML scientists to productionize research, optimize serving, and build robust pipelines with monitoring and observability across the stack at our Amsterdam HQ, with flexible hybrid work options.