Staff Machine Learning Engineer

Written by the T&S Network from Thorn's posting.

Builds machine learning models, classifiers and algorithms for child safety work, aimed at identifying child sexual abuse victims, removing known and new child sexual abuse material from the internet, and detecting abusive content before it spreads. Sits in Thorn's Machine Learning team, where the senior engineer acts as the go-to expert and guides legal, engineering, product, business development and external affairs teams.

• Child sexual abuse material
• Child safety and grooming

What you'd do

• Own end-to-end model development, from design and build through testing and ongoing maintenance.
• Lead ML system designs, set engineering standards, run technical meetings and mentor other engineers.
• Define data sourcing and labeling strategy, and plan future data needs with the Sr. Technical Program Manager.
• Write data labeling guides with specialists, including work with sexually explicit material that is not CSAM.
• Set architectural guardrails and best practices for scalability and reliability across ML systems.
• Speak for Thorn on the technical side at conferences and with outside partners and academic groups.

What it asks for

• 8+ years in machine learning or AI, with 3+ years building production-scale computer vision or NLP systems and pipelines.
• Doctorate or master's in a numeric discipline, or the professional equivalent.
• Working knowledge of Python, PyTorch, TensorFlow with Keras, and scikit-learn.
• Experience with AWS, Terraform, Kubernetes and CI/CD pipelines (preferred).
• Experience with OpenCV, Hugging Face Transformers, FAISS or ONNX runtime (preferred).

Worth knowing

• Involves engaging with sexually explicit material that is not CSAM as part of labeling work.

Terms in this role

CSAM.

Child sexual abuse material: images or video of the sexual abuse of children. Illegal to make, share or possess in nearly every country.

CSEA.

Child sexual exploitation and abuse. The wider harm area around CSAM, including grooming, sextortion and trafficking.

Our best summary of the posting, and it can be wrong. The employer's posting is what counts: read it in full on Thornnorth_east before you apply. If we got something wrong, report this listing.