Machine Learning Engineer, PEDM
Keeper Security- Full Time
- Junior (1 to 2 years)
Candidates should possess 7+ years of experience in Machine Learning, with a focus on representation learning, multimodal modeling, or embedding-based retrieval, and deep domain knowledge in at least one area such as computer vision, natural language processing, or recommendation systems. Strong proficiency in PyTorch is required, including experience fine-tuning foundation models and adapting pretrained vision-language models, along with familiarity with Git, SQL, and Bash.
As a Staff Machine Learning Engineer, Multimodal Modeling, you will lead the advancement of Flock Safety’s core embedding-based retrieval systems, specifically focusing on the scientific aspects of modeling, including fine-tuning and extending multimodal models like CLIP and SigLIP to improve performance and cross-modal alignment. You will unify text and image representations, improve model performance, and ensure extensibility across evolving product use cases, contributing to the delivery of fast, accurate, and scalable search experiences. Additionally, you will customize and extend model architectures and data pipelines to deliver impact, and work on model compression techniques to improve inference efficiency and deployability.
License plate reader cameras for crime prevention
Flock Safety provides a system aimed at enhancing public safety through crime prevention while ensuring privacy and reducing bias. The main product is a network of license plate reader cameras that capture essential vehicle information, which helps in solving crimes. These cameras utilize machine learning technology to ensure that the data collected is objective and ethically used. Flock Safety serves a variety of clients, including neighborhoods, businesses, and law enforcement agencies in over 1,000 cities. The company operates on a subscription model, where clients pay for the installation, maintenance, and access to data and analytics. This approach not only generates a steady revenue stream but also allows clients to benefit from ongoing updates and support. Flock Safety's goal is to create safer communities by providing effective crime prevention tools that respect individual privacy and foster trust between the public and law enforcement.