🎯Work alongside us on innovative projects at the forefront of artificial intelligence. As a research collaborator, you’ll explore a range of impactful topics and contribute to high-quality publications, pushing the boundaries of what's possible in AI.
Topics include:
- Visual Storytelling: Generate textual and/or visual stories (such as images or videos) based on given text and/or visual inputs (like images).
- Talking Face Generation: Create animated faces that can speak and express emotions, used in applications such as virtual news anchors and chatbots, where the digital face gestures and speaks like a real person delivering information or engaging in conversation.
- Prompt-Based Media Generation and Editing: Generate or modify image, audio, and video based on a given prompt.
- Multimodal Logical and Mathematical Reasoning: Understand and solve logical and math problems that integrate text with visual elements like graphs, diagrams, and geometric figures, supporting deeper comprehension in domains such as geometry, set theory, and data analysis.
- Multimodal Causal Reasoning: enhances the system’s ability to understand cause-and-effect relationships across diverse data types, enabling more informed and contextually relevant responses.
- Human-AI Interaction: how people and AI systems communicate and work together, aiming to make AI helpful, understandable, and trustworthy. It includes object detection, segmentation, automatic memory formation, and more.
This collaboration track is open to anyone passionate about research, publication, exploration, or working in a team—whether you're a student, professional, or independent learner. If you're curious and committed, we’d love to collaborate with you. Please note that this research collaboration track is unpaid. Its purpose is to foster teamwork, collective learning, and collaborative problem-solving on complex tasks.
✅ Eligibility & Requirements
- Contributions to open-source or research projects
- Experience in writing research papers
- Proficiency in Python, PyTorch, OpenCV,
- Research publications in reputable conferences or journals
- Familiarity with multimodal reasoning and agent frameworks
- Experience in training or fine-tuning LLMs/multimodal LLMs, including distributed training, LoRA, and quantization
- Strong understanding of deep learning concepts, including overfitting, training instability, and model lifecycle management
- Experience in Vision Transformers, Diffusion Models, YOLO, RT-DETR, Speech/Multimodal Models, and mathematical reasoning models (e.g., Phi-4, DeepSeek-R1, Qwen2.5-Math-7B-Instruct)