Associate ML Engineer, Scalable Recommendations

TikTok USDS Joint Venture; San Jose (CA); Vor Ort; USD 123.000 - 259.000; Vollzeit

Zusammenfassung

TikTok USDS Joint Venture LLC is hiring applied ML engineers and data scientists to advance feed and e-commerce recommendations. You will design and maintain MLOps pipelines for hyper-scale users, architect real-time data pipelines, and optimize CTR/CVR models for production with latency-aware techniques.

You will collaborate with researchers and platform teams to transition experiments into zero-downtime deployments while ensuring robust monitoring and observability of system performance.

Qualifikationen

• Bachelor's degree or higher in Computer Science or related fields.
• Strong programming and problem-solving ability.
• Understanding of MLOps principles, including Docker, model tracking, and workflows.
• Driven to solve complex distributed systems and ML operations challenges.

Aufgaben

• Design, build, and maintain robust MLOps pipelines for large-scale recommender models with CI/CD.
• Architect high-throughput real-time streaming data pipelines and feature stores for dynamic modeling.
• Optimize candidate generation and ranking models for production, including latency reduction with CI/CD and hardware acceleration.
• Build automated monitoring, debugging, and visualization tools for model health, data drift, and resource usage.
• Collaborate with researchers and platform teams to deploy production-scale deep learning models with minimal downtime.

Kenntnisse

Programming

Problem solving

MLOps basics

Distributed systems

Ausbildung

Bachelor's degree in Computer Science or related fields

Tools

Docker

CUDA

TensorRT

Jobbeschreibung

TikTok USDS Joint Venture LLC is hiring applied ML engineers and data scientists to advance feed and e-commerce recommendations. You will design and maintain MLOps pipelines for hyper-scale users, architect real-time data pipelines, and optimize CTR/CVR models for production with latency-aware techniques.

You will collaborate with researchers and platform teams to transition experiments into zero-downtime deployments while ensuring robust monitoring and observability of system performance.