What are the responsibilities and job description for the Intern Research Engineer position at Orbifold AI?
Company Description
Orbifold AI focuses on transforming and enhancing multimodal datasets to create optimized, AI-ready data. By improving the quality and relevance of data, Orbifold enables businesses to build efficient, smart AI models that leverage their unique data for a competitive advantage. The company is dedicated to delivering solutions that save time, reduce costs, and ensure superior outcomes for its clients by offering advanced data transformation and augmentation capabilities. Orbifold is committed to innovation and excellence in the AI space.
Role Description
As a Research Engineer Intern at Orbifold AI, you will be at the forefront of advancing multimodal AI models and infrastructure. Your primary focus will be on developing SOTA visual models, optimizing AI-driven data pipelines, and transforming large-scale, multimodal datasets (text, images, videos) into high-quality training inputs.
This is a full-time, onsite role based in our Los Altos, California office.
- 5 days a week in person (no remote / hybrid).
- Preferred internship duration: 3 months or longer (we strongly prefer candidates who can commit to an extended internship).
We are looking for engineers who are passionate about multimodal AI, large-scale model optimization, and data-centric AI research. You will work with MoE models, distributed systems, and batch inference at an internet scale, pushing the boundaries of training, RAG, and reinforcement learning for enterprise AI applications.
High-performing interns may be considered for a full-time Research Engineer offer upon completion of the internship, based on performance and team needs.
Key Responsibilities
- Develop and optimize multimodal AI models, focusing on computer vision, NLP, and generative architectures.
- Design and maintain high-throughput data pipelines to process and curate large-scale multimodal datasets (images, text, video, structured data).
- Implement and refine MoE models to enhance model efficiency and scalability for multimodal AI training.
- Integrate the latest research innovations into our multimodal AI platform, improving data accuracy, diversity, and relevance.
- Experiment with SOTA models and data curation techniques to maximize AI training efficiency and quality.
- Collaborate with research scientists and machine learning engineers to improve data and AI infrastructure, ensuring scalability as models evolve.
Preferred Qualifications
- Currently pursuing a Bachelor’s, Master’s, or Ph.D. in Computer Science, AI, Machine Learning, or a related field.
- Proficiency in Python and experience working with large-scale open-source datasets like DataComp.
- Strong understanding of distributed computing, HPC, and cloud-based AI infrastructure.
- Hands-on experience with multimodal model training, especially in computer vision, video understanding, and NLP.
- Familiarity with deep learning frameworks and multimodal model architectures.
- Passion for large-scale visual model research, with an ability to work in a fast-paced, dynamic environment.
Location & Internship Details
- Location: Orbifold AI HQ – Los Altos, California
- Work Setup: Onsite, 5 days per week (no remote or hybrid)
- Internship Length: Preferred 3 months (longer commitments are a plus)
- Start Date: Flexible within academic constraints, subject to team needs
Why join Orbifold AI?
- Work on groundbreaking AI research in multimodal model training and enterprise AI.
- Gain hands-on experience with large-scale AI datasets and AI-native applications.
- Collaborate with AI leaders from Google, Meta, and Alibaba to shape the future of enterprise AI.
- Opportunity to contribute to real-world AI innovations with Fortune 500 impact.
- Flexible work culture in a fast-moving AI startup.
If you’re passionate about pushing the boundaries of multimodal AI, we’d love to hear from you!
Apply Now: Send your resume and a short introduction to careers@orbifold.ai.