Research Engineer Generative Video
MirageVisit Mirage website
Mirage is an AI video company building creative tools for video, motion, and sound.
New York City, United States
Funding history
About Mirage
Mirage develops AI-native video-creation products, including the Captions app and Tesseract runtime, alongside foundation models for expressive video generation.
Skills
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will train and optimize large-scale video and multimodal models. You will improve training and inference efficiency, accelerate generation through distillation and quantization, maintain distributed training systems, optimize GPU performance, develop experimentation tooling, and productionize research models.
Requirements
- 2+ years of professional industry experience
- Deep learning systems
- Infrastructure
- PyTorch
- CUDA
- Triton
- Distributed training
- FSDP
- Low-latency inference
- Debugging
- Performance profiling
Responsibilities
- Train and optimize large-scale video and multimodal models
- Improve training and inference efficiency
- Implement distillation, quantization, and pruning techniques
- Build and maintain distributed training systems
- Optimize GPU utilization, parallelism, and throughput
- Develop tooling for experimentation, evaluation, and debugging
- Translate research models into production-ready systems
- Monitor and improve model performance in real-world usage
Benefits
- Comprehensive medical, dental, and vision plans
- 401K with employer match
- Commuter benefits
- Catered lunch multiple days per week
- Dinner stipend for late work
- Grubhub subscription
- Health and wellness perks
- Multiple team offsites per year
- Monthly team events
- Generous PTO policy