AI/ML Engineer

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

Builds the ML systems behind Cinder's trust and safety platform, which platforms use to detect fraud, abuse, and manipulation, enforce policy, and file NCMEC reports. The role sits in the ML team and turns customer moderation decision data into production classifiers and models that shape what content stays up or comes down.

• Fraud and scams
• Coordinated manipulation

What you'd do

• Turn customer decision data into training sets and choose the model approach for each problem.
• Upgrade the classifier pipeline, confidence cascades, and detection methods to trade off cost, speed, and accuracy.
• Create model features that aid moderator decisions, content grouping, and pattern discovery.
• Partner with Engineering on in-house model training, hosting, and inference infrastructure.
• Design the evaluation and metrics infrastructure that classifier scores and model outputs are measured and surfaced through.
• Shape agent evaluation to measure whether AI agents make correct decisions with the right tools and cost.
• Coach teammates and lift the machine learning standard across the company.

What it asks for

• 5 to 8+ years of ML engineering experience on a small team, shipping classifiers and tree-based or embedding models to production.
• Experience taking a classification problem from unlabeled real-world data to a served production model.
• Real experience training classifiers where the target signal is a tiny fraction of the data.
• Working knowledge of LLMs sufficient to judge when they are worth their cost and latency.
• Experience standing up training, serving, evaluation, and monitoring infrastructure from scratch.
• Strong Python and frameworks such as PyTorch, scikit-learn, LangChain, or XGBoost.
• Familiarity with ML deployment practices: automated pipelines, model version control, run tracking, data drift detection, and live monitoring.

Terms in this role

NCMEC.

The National Center for Missing & Exploited Children, a US nonprofit. It runs the CyberTipline, where US-based online services must report apparent child sexual abuse material they find.

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