About the Role

As a Senior ML Engineer at Thorn, this role will build, deploy and maintain world class models, classifiers and algorithms to increase Thorn’s capabilities to achieve our mission of the elimination of child sexual exploitation. They will directly contribute to the Machine Learning team’s critical function in Thorn’s mission - building automated solutions that:

1. Reduce the time it takes to identify a child sexual abuse victim, elevating the most vulnerable and high risk victims from an ocean of online content
2. Enable the removal of child sexual abuse material from the internet, finding known and new content
3. Prevent abuse from happening in the first place, detecting abusive content and behavior before it spreads

They are responsible for ensuring that their work is high quality, as well as collaborating with other ML Engineers to ensure team success. They will thrive in highly ambiguous problem spaces and be able to creatively build scalable solutions with minimum direction. They demonstrate persistence and strong technical expertise to navigate legal, product, and technical constraints. They are responsible for partnering with the engineers, product manager and designers across different products to drive requirements, as well as any relevant external partners, taking into account all parties’ needs.

What You’ll Do

• Own end-to-end model development: design, build, test and maintain machine learning systems and algorithms
• Identify data sourcing and labeling needs for your work, partnering with the Sr. Technical Program Manager to ensure data requirements are visible and prioritized.
• Work closely with specialists at Thorn to create data labeling guides and support labeling efforts. This may involve engaging with sexually explicit material that is not child sexual abuse material
• Review system design/code in close collaboration with other ML Engineers
• Collaborate with engineers on product teams to maintain and deploy models, algorithms, and model serving infrastructure
• Clearly communicate and document problem formulation, feature design, training, and evaluation results to technical and non-technical audiences
• Build relationships and collaborate with external partners
• Present technical work at conferences and other external forums, representing Thorn and sharing insights with the community
• Identify novel solutions for solving online child safety problems. Stay abreast of technical developments and breakthroughs with an eye towards applying solutions to our issue space.
• Collaborate with product managers, designers, and product owners to define scope of work and drive understanding of user needs
• Conduct data analysis as needed to improve cross-functional understanding
• Collaborate with product and engineering in pre-planning/planning sessions, software design, and model development sessions
• Collaborate with UX to engage with users, translating user needs into features and deliverables

What We’re Looking For

• You have a commitment to putting the children we serve at the center of everything you do
• You have a willingness to learn about online child safety and victim identification.
• You have 5+ years background in machine learning and/or artificial intelligence. You also have a Ph.D or master’s degree in a quantitative field, or equivalent professional experience.
• You have experience in designing and building machine learning systems and pipelines for production
• You have experience deploying computer vision and/or natural language processing in production.
• You have an ability and interest in learning new technologies quickly
• You can work with ambiguity, shifting requirements, and collaborate with cross-functional internal and external stakeholders
• You have a passion for machine learning and an aptitude to work in a collaborative environment, can demonstrate empathy for your team and strong advocacy for our users, while balancing the vision and technical constraints of engineering
• You communicate clearly, efficiently, and thoughtfully. We’re a highly-distributed team, so written communication is crucial

Technologies We Use (for relevant roles)

• AWS
• Python
• Pandas
• PyTorch
• scikit-learn/scipy
• Terraform
• CI/CD Pipelines
• Tensorflow with Keras
• Hugging Face Transformers/Diffusers
• OpenCV
• FAISS or other vector stores
• ONNX/onnxruntime
• Kubernetes

Compensation and Benefits

A reasonable estimate of the compensation range for this role is $154,000 - $212,750/year. This range takes into account the wide range of factors that are considered in making compensation decisions, including but not limited to location; skill sets; experience and training; licensure, and certifications.

Our remote-first work model is structured around working from home most of the time. But, there will be times that employees are expected to travel. For example, Thorn may host company-wide gatherings, and smaller teams may hold in-person meetings and team-building events, or require attendance at specific conferences.

At Thorn, we know that great people make a great organization. We value our people and offer employees a broad range of benefits. Learn more about what working at Thorn can mean for you.