Machine Learning Researcher, Internship

A World-Changing Company Palantir creates the world's premier software for data-based decisions and operations. By connecting the right data to the right people, our platforms enable our collaborators to create life-saving medicines, predict supply chain disruptions, locate lost children, and much more. The Role We're in search of research interns to become a part of our Machine Learning Team at Palantir. Our squad focuses on devising innovative machine learning techniques to address our clients' immediate needs and to ensure that Palantir's platforms include unmatched machine learning capabilities. Prospective contenders are graduate students who are enthusiastic about computer vision, natural language processing, and deep learning. If you're driven to expand the limits of what's achievable, and eager to solve challenging problems using unique datasets, this is the perfect opportunity for you! You will assist us in extending the research boundaries across three different areas: Edge AI: Introducing and adapting innovative techniques to facilitate competent object detection model inference on edge hardware varying from ARM and Intel CPU architectures to Jetson NVIDIA GPUs. Investigating both software-only and hardware-informed approaches to optimize the latest object detection architectures on edge hardware. Few-shot learning and vision-language models: Championing novel few-shot learning solutions using advanced vision-language models to enable precise object detection in overhead aerial imagery. Natural Language Processing (NLP): Enabling comprehensive solutions in entity, event, and relationship extraction and classification, as well as in areas of structured text extraction from documents. Technologies We Use Python, PyTorch, ONNX, TensorRT, Hugging Face, Docker. Core Responsibilities Reading, presenting, and implementing research papers in relative research areas of interest. Modifying or inventing techniques in object detection and inference optimization to improve performance on resource-limited hardware like ARM, Intel CPUs, and NVIDIA Jetson GPUs. Fine-tuning vision language models on unique overhead aerial imagery datasets. Modifying or inventing techniques in entity, event, and relationship extraction/classification, and in the area of structured text extraction from documents. What We Value Experience with software-only and/or hardware-informed techniques for model inference optimization on edge devices. A background in Computer Vision Deep Learning, particularly Object Detection. Experience with TensorRT, ONNX, and other optimization frameworks like TVM and OpenVINO. Experience implementing or fine-tuning innovative deep learning models in areas of NER and structured text extraction from documents. An understanding of NLP and deep learning. Familiarity with few-shot learning techniques. Prior experience tuning vision-language models like CLIP or GLIP. Proficiency with any of the primary deep learning frameworks. A bachelor's degree in computer science or related fields. Candidates should be graduate students enrolled in an MS or PhD program. Excellent communication and collaboration skills. Life at Palantir We aim for every Palantirian to achieve their best outcomes, and so we appreciate individuals' strengths, skills, and passions, from your first interview to your long-term growth, rather than relying on traditional career paths. Note: Palantir is dedicated to fostering a culture of diversity, equity, and inclusivity and is delighted to be an Equal Employment Opportunity and Affirmative Action employer. We believe that all Palantirians share the responsibility to uphold our commitment to these principles and welcome applicants with a broad range of backgrounds, perspectives, and life experiences to join us in solving the world's toughest problems. Palantir does not discriminate based on race, religion, color, national origin, gender, sexual orientation, gender identity, gender expression, age, veteran status, disability status, or any other legally protected characteristics.