Graduate ML Engineer — Generative Recommender Research

TikTok; San Jose (CA); On-site; USD 128,000 - 317,000; Full time

Benefits offered by this job

Medical, dental, vision insurance

401(k) with match

Paid parental leave

Disability coverage

Life insurance

Wellbeing benefits

Paid holidays and personal time

Job summary

TikTok is hiring for a graduate-level researcher to join the E-commerce Recommendation Team. You will work on scaling our recommender system with generative approaches, LLMs, and agentic architectures to improve shopping experiences for TikTok users.

You will tackle challenges like cold start, long-tail item distribution, and user intent understanding while advancing training and inference efficiency on GPUs.

Qualifications

• Bachelor’s degree in CS/EE/Math/Statistics or related field.
• Strong foundation in ML and DL with research or project experience in LLMs, RL, generative models, recommender systems, or information retrieval.
• Proficiency in Python and at least one DL framework (PyTorch, TensorFlow, JAX).

Responsibilities

• Drive generative recommender research, exploring generative retrieval and generative re-ranking to improve personalization and training/inference efficiency on GPUs.
• Apply LLMs, RL, and related techniques to enhance semantic understanding and reasoning of recommender systems for e-commerce scenarios.
• Explore self-evolving agents to optimize recommender systems with real-time, personalized decisions.
• Model long-term value and user experience, replacing heuristic systems with advanced LLM/agent capabilities.

Skills

ML & DL fundamentals

Research experience

Strong problem solving

Education

Bachelor’s degree in Computer Science, Electrical Engineering, Mathematics or related

Tools

PyTorch

TensorFlow

JAX

Python

Job description

TikTok is hiring for a graduate-level researcher to join the E-commerce Recommendation Team. You will work on scaling our recommender system with generative approaches, LLMs, and agentic architectures to improve shopping experiences for TikTok users.

You will tackle challenges like cold start, long-tail item distribution, and user intent understanding while advancing training and inference efficiency on GPUs.