18 days ago
Shenzhen, ChinaMid Level
Responsibilities
- Own recommendation and personalization systems end to end, from datasets and training through evaluation, experimentation, rollout, monitoring, retraining, and rollback.
- Build retrieval, ranking, user-modeling, CRM intelligence, and search systems for millions of users.
- Apply classic machine learning, embeddings, fine-tuned open-source models, and LLM-based approaches where they improve results.
- Create evaluation sets and experiment harnesses and run A/B tests, interleaving experiments, and causal analyses.
- Develop batch data, model-serving, and production MLOps workflows with attention to monitoring and GPU costs.
- Collaborate with product, operations, and CRM stakeholders to improve business and product metrics.
- Mentor ML engineers and help raise the team’s recommendation and LLM engineering practices.
Requirements
- At least 2 years of industrial machine learning experience, typically including shipped recommendation, personalization, ranking, or search systems serving millions of users.
- Strong knowledge of retrieval and ranking modeling, embeddings, and online experimentation.
- Hands-on experience fine-tuning open-source models and LLMs using SFT, LoRA, or DPO for ranking, personalization, or user modeling.
- Experience building evaluation sets and model evaluation harnesses.
- Understanding of dataset licensing and provenance for commercial use.
- Solid MLOps knowledge covering data pipelines, productionization, monitoring, and GPU cost awareness.
- Experience with batch data stacks, schedulers, and cloud training and serving platforms.
- Strong Python and PyTorch skills.
- Familiarity with Hugging Face, transformers, PEFT, TRL, and modern inference stacks such as vLLM is a plus.
- Education in a quantitative field such as Computer Science, Statistics, or Mathematics, or equivalent practical depth.
- Ability to solve complex problems with scientific rigor, work through ambiguity, and iterate quickly.
- Experience using agentic AI tools in daily engineering work and demonstrating their workflow impact.
Benefits
- Career growth opportunities with increasing challenges and responsibilities.
- Opportunity to join a high-growth team scaling globally.
- Competitive performance-based compensation.
- Candid, open, and collaborative culture with valued feedback.
Tech Stack
Categories
