AI Research Lead

📋 Description

• • Perplexity is at the forefront of revolutionizing how people access and interact with information, and at the heart of this mission are our proprietary Online Large Language Models (LLMs), the Sonar models. We are seeking an exceptional AI Research Tech Lead to spearhead our research strategy and guide the development of these cutting-edge models. This is a unique opportunity to shape the future of AI-powered information retrieval and to lead a talented team of researchers in pushing the boundaries of what's possible.
• • As the AI Research Tech Lead, you will be instrumental in defining and executing the macro research direction across various modalities. This includes not only advancing our post-training LLM capabilities for agent trajectories but also charting the course for future mid-training initiatives. Your strategic vision will be crucial in developing a research roadmap that continuously enhances the performance of our Sonar models, ensuring we deliver a state-of-the-art (SOTA) Online LLM experience to our global user base. You will leverage our incredibly rich and unique query-answer dataset, a significant asset, to scale our model performance and maintain our competitive edge.
• • Research Leadership and Strategy:
• • Define and execute the overarching research direction for Perplexity's AI initiatives, with a particular focus on multiple modalities and advanced LLM techniques.
• • Lead the strategic planning and development of a comprehensive research roadmap designed to significantly advance the capabilities of our Sonar models.
• • Drive innovation in supervised and reinforcement learning techniques specifically tailored for query answering tasks, aiming for unprecedented accuracy and relevance.
• • Collaborate closely with Perplexity's leadership team to ensure research priorities are tightly aligned with overarching product goals and business objectives, maximizing the impact of our research efforts.
• • Foster a culture of continuous exploration and experimentation, encouraging the team to pursue novel approaches and challenge existing paradigms in AI research.
• • Team Development and Mentorship:
• • Coach, mentor, and inspire a high-performing team of AI research scientists and engineers, nurturing their technical expertise and professional development.
• • Establish and communicate a clear, long-term macro research direction for the team, guiding our progress across different modalities and model architectures.
• • Lead the end-to-end hiring process for new research talent, identifying and attracting individuals who will contribute to our innovative environment.
• • Cultivate a deeply collaborative and supportive team environment that actively encourages knowledge sharing, open discussion, and collective problem-solving.
• • Empower team members to take ownership of their projects, providing them with the autonomy and resources needed to succeed.
• • Technical Excellence:
• • Spearhead the post-training of SOTA LLMs for query answering, employing the most advanced supervised and reinforcement learning techniques available.
• • Take full ownership of and meticulously optimize the entire stack of data, training, and evaluation pipelines essential for cutting-edge LLM post-training.
• • Deliver Sonar models that consistently achieve SOTA performance in query answering, setting new benchmarks for accuracy, speed, and user satisfaction.
• • Lead in-depth research into sophisticated agent trajectories and multi-modal capabilities, expanding the versatility and intelligence of our models.
• • Drive the technical roadmap for future investments in mid-training and potentially pre-training initiatives, ensuring Perplexity remains at the vanguard of LLM development.
• • Stay abreast of the latest advancements in AI and Deep Learning, proactively integrating new methodologies and technologies into our research and development processes.
• • Cross-Functional Collaboration:
• • Work in seamless partnership with our dedicated engineering teams to ensure the efficient and effective integration of Sonar models into Perplexity's product ecosystem.
• • Collaborate closely with product management teams to gain a deep understanding of user needs and translate these insights into actionable research priorities.
• • Partner with our data science and engineering teams to strategically leverage our unique and extensive query-answer dataset, unlocking its full potential for model improvement.
• • Clearly and effectively communicate research progress, key findings, and strategic implications to stakeholders across all levels of the organization, ensuring alignment and buy-in.
• • Contribute to Perplexity's overall technical strategy, offering insights and expertise that inform product development and long-term vision.