Alice is integrated into a wide range of smart devices, including speakers, smartphones, TVs, smart home appliances, computers, and laptops. It also powers Yandex products such as Search, Maps, Browser, and Smart Assistant. To enable intelligent conversations, we build and deploy advanced ML models that understand context, predict intent, and generate human-like responses.
We are looking for an AI Consultant to help advance LLM architectures, deliver production-ready models, and improve the user experience across Alice interfaces.
Tasks that await you
Build and align LLMs
You will architect, train, and optimize large-scale LLMs for dialogue systems, from pretraining to RLHF/DPO alignment and deployment.
Advance multilingual and fast inference
You will develop methods for multilingual understanding and low-latency inference.
Boost Alice with model deployment
You will deploy models in Alice's ecosystem to improve metrics such as engagement, task success, and user satisfaction.
Scale MLOps with massive data
You will design and maintain MLOps pipelines for continuous improvement, leveraging petabyte-scale interaction data.
We expect that you
- Have shipped production LLMs with expertise in at least one of the following: pretraining 100B+ parameter models, alignment techniques (RLHF, DPO, reward modeling), or reinforcement learning for dialogue systems
- Have mastered the Python/PyTorch stack, distributed training, and ML system design
- Balance innovation with production constraints (latency, scalability)
- Critically evaluate model performance using metrics and user feedback
- Are fluent in English (reading and writing) and have HSK4-level proficiency in Chinese
More about our team
Why Alice and Yandex?
- Shape the future of AI for millions of users
- Solve unique, high-impact technical challenges at scale
- Collaborate with industry-leading ML engineers and scientists