About the job
About SimpliSafe
SimpliSafe stands at the forefront of innovation in the home security sector, committed to ensuring that every household feels secure. Our mission is to deliver accessible and comprehensive security solutions through user-centric designs that empower individuals and families to protect their most valued assets.
We foster a collaborative and dynamic work environment where continuous learning and professional growth are paramount. Our teams consist of passionate and talented individuals dedicated to technology, security, and providing outstanding customer experiences.
Embracing a hybrid work model, we encourage our teams to balance their time between office collaboration and remote work. Typically, our teams gather in our state-of-the-art office on core days—usually Tuesday, Wednesday, or Thursday—allowing for personal workspace flexibility while maximizing in-person collaboration.
Why Join Us?
As we expand and excel, we are on the lookout for intelligent, talented, and humble individuals who align with our values, eager to innovate in the home security landscape and relentlessly pursue our mission of securing Every Home.
About the Role
We are in search of a driven and highly skilled Computer Vision Applied Research Engineer to become a pivotal member of our expanding Edge AI team. In this influential role, you will spearhead the development of on-device machine learning solutions for outdoor monitoring in the home security domain. Your expertise will involve building and optimizing computer vision models that operate in real-time on resource-constrained embedded devices such as outdoor and doorbell cameras, striking the right balance between accuracy, latency, memory usage, power efficiency, and reliability under challenging conditions (including low light, adverse weather, and motion blur).
Key Responsibilities:
- Lead the comprehensive development of edge ML models for outdoor monitoring applications (e.g., detecting and classifying persons, vehicles, and packages; tracking; segmentation; and event understanding).
- Design, train, and deploy transformer-based vision models (e.g., compact ViTs, hierarchical transformers, DETR-style detectors) and hybrid CNN-transformer architectures optimized for embedded inference.
- Enhance model efficiency through resource-aware design and training approaches, including:
- Architecture: Token/patch reduction, efficient attention variants, early-exit/conditional compute.
- Training: Distillation from large transformer models to edge implementations.
- Compression: Techniques such as Quantization (PTQ/QAT), pruning, mixed precision, and operator-aware optimization.
- Collaborate closely with cross-functional teams to integrate models into our products and ensure they meet quality and performance standards.

