注册一亩三分地论坛,查看更多干货!
您需要 登录 才可以下载或查看附件。没有帐号?注册账号 
x
Job Title: Research Scientist for Multimodal Large Model Development
Location: Northen Virginia or remote
Job Type: Full-time
About SensForge:
SensForge is a dynamic and fast-growing company specializing in smart home camera technology. We currently have multiple openings focusing on multimodal large model development in our R&D team at different levels, including Research Scientists, Senior research Scientist, and Principal research scientist.
Key Responsibilities:
This position is suitable for candidates who are passionate about training, finetuning, and optimizing multimodal large models (especially vision language models) that can be deployed at edge or on device for smart home applications. The job responsibilities of this position include: - Build the data processing pipeline to prepare the required data for model development
- Adjust multimodal large model architecture, loss function, and training recipes to meet the model requirements for the target use cases
- Train and/or finetune multimodal large models and monitor training processes and adjust training hyperparameters according to the key performance indicators
- Evaluate the model performance using off-the-shelf model evaluation toolkits and analyze the evaluation results and determine the next steps for model improvement
- Prune and quantize the model to reduce the model memory footprint and improve the model throughput while maintaining the model accuracy
- Deploy the model for production using popular the model deployment toolkits
- Set up multimodal large model development environment as a docker container and/or Python virtual environment and install the required system and Python packages
Required Qualifications: - Bachelor's degree or above
- Hands-on experience in deep model development in the area of computer vision
- Familiar with the theory and practices of modern deep learning, including CNN and transformer-based models in the area of computer vision and natural language processing
- Familiar with prompting large language models and multimodal large models
- Familiar with PyTorch, Python, Bash, Docker and Git
- Strong problem-solving skills and excellent communication skills
- Good team player and demonstrated anti-pressure abilities
Desired Qualifications: - Demonstrated experience in large model development
- Expert in large language models and multimodal large language models such as vision language models
- Familiar with the latest advances and landscape of vision language models
- Familiar with the toolkits for multimodal large model evaluation and deployment
- Expert in prompt engineering
- Demonstrated leadership capabilities
- Bilingual in both English and Chinese
How to Apply:
To apply, please send your resume and cover letter to developer@sensforge.com |