About this role
Summary: Interns will develop and improve video-language foundation models for fine-grained video understanding and robotics applications. Responsibilities include building ML pipelines, creating evaluation benchmarks, and exploring multimodal learning and agentic AI techniques.
### Job Description:
About the Team
We are building Max VLM, a next-generation video foundation model designed to understand human activities, object interactions, and the physical world through first-person video.
Our team focuses on developing state-of-the-art video understanding systems and applying them to real-world applications, including fine-grained video understanding, robotics evaluation, and intelligent video analysis.
As a fast-moving startup, we combine cutting-edge AI research with practical engineering, working across the entire machine learning lifecycle, from data generation and model training to evaluation, optimization, and deployment.
We are looking for motivated interns who are passionate about multimodal AI and eager to tackle challenging real-world machine learning problems.
Responsibilities
1. Develop and improve video-language foundation models for fine-grained video understanding, temporal reasoning, and human-object interaction analysis.
2. Build and optimize machine learning pipelines for data processing, model training, evaluation, and inference.
3. Develop evaluation methods and benchmarks for video understanding and robotics applications, including action understanding, manipulation tasks, and task-success assessment.
4. Explore advanced techniques in multimodal learning, reinforcement learning, and agentic Al to improve model capabilities.
5. Collaborate with the team to translate new ideas into practical solutions and contribute to research and product development.
Qualifications
Minimum Qualifications
- Currently pursuing a Master's or PhD degree in Computer Science, Artificial Intelligence, Machine Learning, Computer Vision, Robotics, or a related field. Solid understanding of deep learning, with experience in at least one of the following areas: computer vision, multimodal learning, large language models, video understanding, or robotics.
- Proficiency in Python and experience with deep learning frameworks such as PyTorch.
- Strong problem-solving skills and the ability to independently implement and evaluate machine learning algorithms.
Preferred Qualifications
- Hands-on experience with vision-language models, video understanding, or multimodal foundation models.
- Experience with LLM/VLM fine-tuning, reinforcement learning, or agentic AI.
- Experience with robotics, robotic manipulation, or robot learning and evaluation.
- Familiarity with distributed training, model optimization, or inference frameworks such as DeepSpeed, Megatron-LM, or vLLM.
- Publications at leading Al, computer vision, or robotics conferences, such as NeurIPS, ICML, ICLR, CVPR, ICCV, RSS, or CoRL, are a strong plus.
- Passion for applying cutting-edge Al technologies to challenging real-world problems.
Research & Growth Opportunities
- Real-World Impact: Work on practical AI challenges with real-world video data and direct applications in video intelligence and robotics.
- Hands-on Experience: Gain experience in developing, training, evaluating, and optimizing state-of-the-art multimodal models.
- Frontier Innovation: Explore emerging techniques in video reasoning, agentic Al, and reinforcement learning.
- Research & Publications: Opportunities to contribute to novel research and high-quality publications at leading AL, computer vision, and robotics conferences.
- Startup Environment: Work closely with an experienced team, take ownership of meaningful projects, and see your contributions translate into real-world impact.
Default Benefits:
- Health insurance
- Vision care
- Dental coverage
- 401(k)
- Paid holidays
- PTO (Paid Time Off)
- Sick leave