Senior Machine Learning Engineer, GenAI Security

Senior Machine Learning Engineer, GenAI Security

Remote - United States

Reddit is a community of communities. It’s built on shared interests, passion, and trust, and is home to the most open and authentic conversations on the internet. Every day, Reddit users submit, vote, and comment on the topics they care most about. With 100,000+ active communities and approximately 126 million daily active unique visitors, Reddit is one of the internet’s largest sources of information.

For more information, visit

The GenAI Security team within Reddit’s Security, Privacy, Assurance, and Corporate Engineering organization protects Reddit’s GenAI usage across employee tools, internal agents, and production user-facing systems. Our mission is to secure and protect Reddit’s AI traffic and GenAI adoption by default.

We are building zero-trust, defense-in-depth systems that verify identity, permissions, data access, and semantic intent across AI workflows. A core part of this work is developing practical, high-quality ML models that detect and prevent security risks such as prompt injection, jailbreak attempts, sensitive data exfiltration, unsafe model behavior, anomalous usage, and unauthorized agent actions.

We are looking for a Senior Machine Learning Engineer to lead model development for GenAI Security and help establish strong ML practices across SPACE. This role owns the full machine learning lifecycle: problem definition, data ETL, feature engineering, model training, model evaluation, deployment, experimentation, prediction, monitoring, debugging, and retraining.

• Build and improve security-focused ML models for Reddit’s GenAI traffic, including guardrail models, semantic classifiers, anomaly detection models, and other neural network based security signals.
• Own model development end to end: define the security problem, assemble and label datasets, build ETL pipelines, engineer features, train models, evaluate quality, deploy to production, monitor performance, and retrain from production feedback.
• Use modern deep learning architectures, including neural networks, transformers, sequence models, embeddings, and model distillation where they are the right practical fit.
• Design rigorous evaluation suites for adversarial examples, hard negatives, long-context inputs, structured payloads, tool calls, multi-turn workflows, and real production traffic.
• Improve model precision, recall, latency, cost, calibration, and operational reliability for high-impact production surfaces.
• Build repeatable MLOps workflows for SPACE, including training pipelines, model lineage, artifact management, holdout evaluation, dashboards, rollback paths, and retraining loops.
• Partner closely with ML Infrastructure, LLM Gateway, DevX, Ads, Answers, Safety, Privacy, Compliance, and other Security teams to bring security models into real production workflows.
• Work pragmatically with Reddit’s evolving ML platform, using existing infrastructure where possible and building focused tooling when needed to keep model iteration moving.
• Translate security goals into measurable model outcomes and help partners understand tradeoffs between risk reduction, latency, false positives, and product impact.
• Provide technical direction to other engineers and serve as a go-to ML expert for GenAI Security and broader SPACE model needs.

• 5+ years of experience building, training, evaluating, and deploying production ML or deep learning models.
• Hands‑on experience with modern ML frameworks such as PyTorch, Tensor Flow, or similar.
• Strong practical understanding of the full ML lifecycle: problem definition, data ETL, feature engineering, training, evaluation, deployment, monitoring, debugging, and retraining.
• Experience building data pipelines and working with large‑scale datasets.
• Experience designing rigorous model evaluations, including precision/recall/F1, false positive analysis, threshold tuning, calibration, holdout sets, regression tests, and production‑quality validation.
• Experience shipping production‑quality software, preferably in Python and/or Go.
• Strong communication skills and ability to explain model behavior, risk tradeoffs, and…