Internship - Machine Learning Research Engineer

Perplexity AI, Inc.

Job Overview

Location

Remote

Job Type

Full-time

Full Job Description

📋 Description

• Embark on a transformative 12-24 week full-time internship at Perplexity, a company at the forefront of revolutionizing information discovery. This is an unparalleled opportunity to immerse yourself in the dynamic world of Machine Learning Research Engineering, contributing directly to the advancement of search quality. You will be an integral part of our mission to push the boundaries of what's possible in information retrieval and answer generation, working alongside a team of world-class researchers and engineers.
• Your primary objective will be to relentlessly drive improvements in search quality. This isn't limited to a single approach; you will have the autonomy to explore and implement solutions through innovative models, sophisticated data strategies, advanced tooling, or any other leverage you can identify and develop. This role demands a proactive and experimental mindset, encouraging you to think outside the box and challenge existing paradigms.
• A significant aspect of your work will involve the training and optimization of large-scale deep learning models. You will gain hands-on experience with cutting-edge frameworks such as PyTorch, a cornerstone of modern deep learning research. The scale of our models necessitates proficiency in distributed training techniques. You will learn to leverage distributed training libraries like PyTorch Distributed, DeepSpeed, or Fully Sharded Data Parallel (FSDP) to efficiently train massive neural networks across multiple compute nodes.
• Furthermore, you will become adept at harnessing hardware acceleration, understanding how to optimize model performance on specialized hardware like GPUs and TPUs. The focus of your model development will be on retrieval and ranking models, which are critical components of any effective search engine. You will delve into the intricacies of how to best represent and rank information to provide users with the most relevant results.
• Beyond model training, you will engage in cutting-edge research within the field of representation learning. This includes exploring and implementing advanced techniques such as contrastive learning, a powerful method for learning representations from unlabeled data. You will also contribute to research in multilingual modeling, ensuring our search capabilities are effective across diverse languages. Developing robust evaluation methodologies for these models will be crucial, as will exploring multimodal modeling, which integrates information from various sources like text, images, and audio to enhance search relevance.
• A key area of focus will be the development and optimization of Retrieval-Augmented Generation (RAG) pipelines. These pipelines are essential for grounding generated answers in factual information, ensuring accuracy and reliability. You will work on building efficient and effective RAG systems that seamlessly integrate retrieval mechanisms with generative models to produce high-quality, contextually relevant answers.
• Throughout your internship, you will be exposed to the full research lifecycle, from ideation and experimentation to implementation and evaluation. You will have the opportunity to present your findings and contribute to the broader research community. This role is ideal for individuals who are passionate about making a tangible impact on how people access and interact with information, and who are eager to learn and grow in a fast-paced, research-intensive environment.
• You will collaborate closely with senior researchers and engineers, benefiting from their mentorship and expertise. This collaborative environment fosters learning and encourages the sharing of knowledge and best practices. Your contributions will be valued, and you will have the chance to see your work directly influence the product and its users. This internship is more than just a learning experience; it's an opportunity to be part of something truly innovative and to shape the future of search.

🎯 Requirements

• Strong foundational understanding of search and retrieval systems, including principles of information retrieval, ranking algorithms, and evaluation metrics.
• Demonstrated proficiency in PyTorch, with practical experience in training and optimizing deep learning models, ideally including experience with distributed training frameworks (e.g., PyTorch Distributed, DeepSpeed, FSDP) and performance optimization techniques for large-scale models.
• Genuine interest and theoretical understanding of representation learning, encompassing areas such as contrastive learning, dense and sparse vector representations, representation fusion, cross-lingual representation alignment, training data optimization, and robust evaluation methodologies.
• A strong academic background with a publication record in reputable AI/ML conferences or workshops (e.g., NeurIPS, ICML, ICLR, ACL, EMNLP, SIGIR) is highly desirable, showcasing a commitment to research and contribution to the field.

🏖️ Benefits

• Gain invaluable hands-on experience in a cutting-edge AI research environment, working on real-world problems with significant impact.
• Mentorship from world-class researchers and engineers, providing guidance and opportunities for professional growth.
• Opportunity to contribute to a product used by millions, directly influencing the future of information discovery.
• Exposure to state-of-the-art machine learning techniques, distributed systems, and large-scale model training.
• A dynamic and collaborative work environment that fosters innovation and learning.

Skills & Technologies

PyTorch

Junior

Remote

Ready to Apply?

Apply Externally

You will be redirected to an external site to apply.

Perplexity AI, Inc.

Visit Website

About Perplexity AI, Inc.

Perplexity AI operates an AI-powered conversational search engine that answers queries by synthesizing live web information. The platform combines large language models with real-time retrieval, citing sources for transparency. Founded in 2022, the San Francisco-based company offers free and subscription tiers, mobile apps, and browser extensions, targeting consumers and enterprises seeking accurate, verifiable answers instead of traditional link lists.

View Company Profile