Staff Research Engineer, Neural Architectures

Ephapsys; San Francisco (CA); On-site; USD 180,000 - 280,000; Full time

Benefits offered by this job

Early-stage equity participation

High research autonomy

Job summary

Ephapsys in San Francisco seeks a Staff Research Engineer, Neural Architectures to advance ephaptic coupling in modern AI systems. You will work with the founder to formalize models, design experiments, and achieve measurable breakthroughs in performance, efficiency, and robustness.

This role emphasizes producing publishable research and translating findings into production-ready Ephapsys SDK components with platform integration and benchmarking.

Qualifications

• PhD in Machine Learning, AI, Applied Mathematics, Computational Neuroscience, or related field.
• Strong mathematical foundation (linear algebra, optimization, dynamical systems).
• Deep expertise in neural network theory and architecture design.
• Advanced proficiency in PyTorch or TensorFlow.
• Experience conducting reproducible research with controlled experimental design.

Responsibilities

• Leading the design and formalization of new ephaptic interaction operators in deep learning.
• Developing mathematical models describing cross-neuron interaction dynamics.
• Integrating ephaptic mechanisms into transformers, MLPs, and hybrid architectures.
• Conducting ablation studies and theoretical analysis of convergence and stability.
• Translating research into modular, production-ready SDK components.

Skills

Neural network theory

Mathematical foundations

Optimization

Dynamical systems

Research execution

Education

PhD in ML/AI/Applied Math/Computational Neuroscience

Tools

PyTorch

TensorFlow

Job description

About Ephapsys

We leverage ephaptic coupling to pioneer a new paradigm for empowering and governing autonomous AI agents across industries.

As billions of agents spread across clouds, devices, and biological substrates, Ephapsys aims to be the decentralized identity and integrity layer that keeps every agent trusted, safe, and under control.

The Role

We are seeking a Staff Research Engineer, Neural Architectures to lead the advancement of ephaptic coupling technology across modern AI systems.

This role is research-intensive and focused on generating novel, experimentally validated improvements in neural architecture design.

You will work directly with the founder to formalize theoretical models, design controlled experiments, and deliver measurable breakthroughs in performance, computational efficiency, robustness, and adaptability.

The role includes producing publishable research while ensuring practical integration into the Ephapsys SDK and platform.

Responsibilities

• Neural Architecture Innovation
• Design and formalize new ephaptic interaction operators within deep learning architectures
• Develop mathematical models describing cross-neuron interaction dynamics
• Integrate ephaptic mechanisms into transformers, MLPs, and hybrid architectures
• Conduct rigorous ablation studies and theoretical analysis of convergence behavior
• Investigate stability properties and dynamical systems implications of modulation layers

Performance & Efficiency Breakthroughs

• Deliver measurable improvements in training efficiency, inference latency, or generalization
• Explore adaptive modulation strategies for improved robustness and transfer learning
• Analyze computational complexity and scalability trade-offs
• Evaluate hardware-aware and distributed training optimizations

Research & Publication

• Produce internal technical reports documenting experimental findings
• Contribute to peer-reviewed publications and conference submissions
• Present research findings in technical forums when appropriate
• Stay current with frontier model research including transformer alternatives and emerging paradigms

SDK & Platform Integration

• Translate validated research into modular, production-ready SDK components
• Build reusable frontier model modules and research-grade agents
• Establish benchmarking standards to quantify ephaptic performance gains
• Collaborate with security engineering to evaluate new attack surfaces introduced by novel operators

Qualifications

Required

• PhD in Machine Learning, AI, Applied Mathematics, Computational Neuroscience, or related field
• Strong mathematical foundation (linear algebra, optimization, dynamical systems)
• Deep expertise in neural network theory and architecture design
• Advanced proficiency in PyTorch or TensorFlow
• Experience conducting reproducible research with controlled experimental design
• Demonstrated ability to translate theory into working implementations

Preferred

• Experience training or fine-tuning large-scale models
• Familiarity with distributed training and performance optimization
• Prior publications in reputable AI/ML conferences or journals
• Interest in expanding neural computation beyond synaptic-only paradigms
• Experience contributing to open-source ML frameworks

What Success Looks Like

Within the first 3-6 months, you will have demonstrated experimentally validated performance or efficiency gains through ephaptic integration, contributed to at least one publishable research milestone, and delivered production-ready frontier model capabilities integrated into the Ephapsys platform.

Why Join Ephapsys

• Shape the future of neural architecture research
• Operate at the frontier of adaptive and secure AI systems
• High research autonomy and technical ownership
• Early-stage equity participation
• Direct collaboration on foundational AI innovation