Research Scientist

**AI Research Scientist (Test Time Compute) | **naptha.ai**:
**About the role**:
We are seeking an exceptional AI Research Scientist to join Naptha AI at the ground floor, focusing on advancing the state of the art in test time compute optimization for large language models. In this role, you will be responsible for researching and developing novel approaches to improve inference efficiency, reduce computational requirements, and enhance model performance at deployment. Working directly with our technical team, you will help shape the fundamental architecture of our inference optimization platform.

This role is critical in solving core technical challenges around model compression, efficient inference strategies, and deployment optimization. You will work at the intersection of machine learning, systems optimization, and hardware acceleration to develop practical solutions for real-world model deployment and scaling.

**Core Responsibilities**:
**Research & Development**:

- Design and implement novel architectures for efficient model inference
- Develop frameworks for model compression and quantization
- Research approaches to optimize test-time computation across different hardware
- Create efficient protocols for distributed inference and resource management
- Implement and test new ideas through rapid prototyping

**Technical Innovation**:

- Stay at the forefront of developments in ML efficiency and inference optimization
- Identify and solve key technical challenges in model deployment
- Develop novel approaches to model compression and acceleration
- Bridge theoretical research with practical implementation
- Contribute to the academic community through publications and open source

**Platform Development**:

- Help design and implement efficient inference pipelines
- Develop scalable solutions for model deployment and serving
- Create tools and frameworks for performance monitoring and optimization
- Collaborate with engineering team on implementation
- Build proofs of concept for new optimization techniques

**Leadership & Collaboration**:

- Work closely with engineering team to implement research findings
- Mentor team members on advanced optimization techniques
- Contribute to technical strategy and roadmap
- Collaborate with external research partners when appropriate
- Help evaluate and integrate external research developments

**In this role, you're a good fit if you have**:

- Strong background in machine learning and systems optimization
- Deep understanding of model compression and efficient inference techniques
- Hands-on experience with modern ML frameworks and deployment tools
- Experience with ML infrastructure and hardware acceleration
- Track record of implementing efficient ML systems
- Excellent programming skills (Python required, C++/CUDA a plus)
- Strong analytical and problem-solving abilities
- PhD in Machine Learning, Computer Science, Mathematics, or equivalent experience is a plus
- Published research in relevant fields is a plus

**Required Technical Experience**:

- Python programming and ML frameworks (PyTorch, TensorFlow)
- Experience with model optimization techniques (quantization, pruning, distillation)
- MLOps and efficient model deployment
- Hardware acceleration (GPU, TPU optimization)
- Version control and collaborative development
- Experience with large language models

**About the hiring process**:

- Initial technical interview
- Research presentation
- System design discussion
- Technical challenge
- Team collaboration interview

**Compensation & Benefits**:

- Competitive salary with significant equity stake
- Remote-first work environment
- Full medical, dental, and vision coverage
- Flexible PTO policy
- Learning and development budget
- Conference and research publication support
- Home office setup allowance

**Additional Notes**:

- Must be comfortable with ambiguity and rapid iteration typical of pre-seed startups
- Strong bias for practical implementation of research ideas
- Passion for advancing the field of efficient ML systems
- Interest in open source contribution and community engagement