Perplexity14.04.2026
Member of Technical Staff (AI Researcher)
Полная занятостьSan Francisco
Обязанности
- 01Post-train SOTA LLMs using the latest supervised and reinforcement learning techniques (SFT/DPO/GRPO)
- 02Leverage our rich query/answer dataset to scale model performance across Sonar, Deep Research, Comet, and Search products
- 03Stay current with the latest LLM research, especially in model training, optimization, and personalization techniques
- 04Implement preference optimization and personalization capabilities to enhance user experience
- 05Invent in-house improvements and optimizations to enhance SOTA models
- 06Turn research ideas into algorithms and run experiments to launch new models
- 07Own full-stack data, training, and evaluation pipelines required for model development
- 08Build robust and effective training frameworks (on top of Megatron/PyTorch) for post-training LLMs
- 09Implement necessary infrastructure and components to support cutting-edge model training at scale
- 10Integrate models seamlessly into our product ecosystem
- 11Work closely with engineering teams to integrate models into Perplexity's product suite
- 12Collaborate across teams to ensure cohesive AI experiences throughout our platform
- 13Partner with product teams to understand user needs and translate them into model improvements
Требования
- 01Proven experience with large-scale LLMs and Deep Learning systems
- 02Strong programming skills in Python/PyTorch; versatility is a plus
- 03Experience with post-training techniques and reinforcement learning
- 04Self-starter with a willingness to take ownership of tasks
- 05Passion for tackling challenging problems
- 06Minimum 2-6 years of experience on relevant projects (depending on seniority level)
- 07PhD in Machine Learning, AI, Systems, or related areas
- 08Experience in post-training LLMs with SFT/DPO/GRPO
- 09C++/CUDA programming skills
- 10Experience building LLM training frameworks
- 11Academic publications and research impact
- 12Experience with agent systems and multi-step reasoning
- 13Background in personalization and preference learning