AI role · Open offer
Senior ML Engineer — recommendation & LLM
ByteDance
This listing is optimized for AI-native talent: clear skills and tools so candidates know if they fit before they apply—and so you spend less time screening mismatches.
4 skills2 tools
Share this role
At a glance
How you'll work, where, and key listing dates.
- Employment
- Full-time
- Workplace
- Hybrid
- Location
- Zhichun Road, Haidian District, Beijing 100086, China · Hybrid
- Compensation
- CNY 450k–750k/year + bonus
- Experience level
- Senior
- Team
- Douyin & TikTok ML
- Benefits & perks
About this role
What the team is looking for and how this position fits.
ByteDance is hiring a Senior Senior ML Engineer — recommendation & LLM on the Douyin & TikTok ML team (Hybrid · Beijing, China). This is a hands-on role in LLM & generative AI—not a generic “AI enthusiast” post.
You will own LLM features, retrieval, agents, and evaluation end to end: problem framing, offline evaluation, safe rollout, and iteration from production signals. Success is measured with clear metrics (quality, latency, cost, or business KPIs)—not slide decks alone.
**ByteDance** builds **TikTok**, **Douyin**, and global content platforms — recommendation and generative AI at billion-user scale from **Beijing Haidian**.
Senior ML Engineer (hybrid Beijing): **feed ranking**, **multimodal understanding**, and **LLM features** with strict latency budgets.
About ByteDance
- Global technology company; Haidian R&D center for core ML systems.
- Real-time recommendation infra handles massive QPS.
- Mandarin-primary teams with international research reading.
What you will do
- Improve ranking models with causal and long-term engagement metrics.
- Build multimodal embeddings for video and text content.
- Prototype LLM assistants with content-safety filters.
Stack & environment
- Python, PyTorch, Spark, Kubernetes.
- Hybrid Beijing Haidian.
What we look for
- Large-scale recommender or NLP systems.
- Eligible to work in China; Mandarin proficiency.
Skills we're hiring for
What each capability means for this role—not a keyword checklist.
- Python
Production Python for data pipelines, model code, APIs, and automated tests—not notebook-only workflows.
- PyTorch
Training and inference with PyTorch: reproducible experiments, checkpointing, and deployment-minded exports.
- NLP
Tools & stack
Platforms and products mentioned for this role (APIs, IDEs, vector DBs, CRMs, design tools, etc.).
- KubernetesTool
Deploying and scaling model services with health checks, autoscaling, and rollout discipline.
- SparkTool
Distributed feature generation and batch scoring on large tabular or event datasets.
You're a strong fit if…
Tailored to this listing—use it as a self-check before you apply.
- ✓You can show evidence in your application: Feed ranking, multimodal content understanding, or LLM features.
- ✓Several of these show up in recent shipped work—not only on your CV: Python, PyTorch, NLP, SQL, and Kubernetes.
- ✓You have owned LLM features, retrieval, agents, and evaluation in production: debugging live issues, running postmortems, and iterating from real user or business signals.
- ✓You are comfortable with the hybrid rhythm (Hybrid · Beijing, China)—onsite collaboration when it matters, deep work when remote.
- ✓You explain trade-offs (quality, cost, latency, safety) in plain language—with numbers or examples, not buzzwords.
- ✓
Apply
Candidate accounts only: add your headline and bio on , and optionally a CV link or PDF so recruiters see your full résumé. Then write a short pitch below—applications use your login email.
What the employer asked for
Feed ranking, multimodal content understanding, or LLM features.
Sign in with a candidate account to apply
Register, add a headline and bio, then you can attach a CV link or PDF on My account before you apply. That keeps applications tied to your profile and makes it easier for employers to review your background.
Suggested
AI talent you may also like
Profiles ranked by overlap with this role's skills and tools—handy if you're hiring for a team or comparing backup candidates.
No close matches in the directory yet—try with your own filters.
Similar open roles
Other listings that share skills or tools with this one—useful if you want comparable stacks or backup options.
·