AI Content Moderation Engineer

About the Role

We are looking for an experienced AI Content Moderation Engineer to design, train, deploy, and maintain automated content moderation systems for a social media platform. The system will automatically analyze uploaded videos, livestreams, images, audio, captions, comments, and user-generated content to identify and enforce policies against inappropriate or harmful content.

The ideal candidate will have strong expertise in machine learning, deep learning, computer vision, natural language processing, audio analysis, and real-time AI inference. The goal is to build a scalable AI-powered moderation system that operates automatically with minimal human intervention.

Key Responsibilities

1. AI Model Development and Training

• Design, train, fine-tune, and deploy AI models for automated content moderation.
• Develop video classification models to detect sexual content, nudity, explicit sexual activity, violence, graphic imagery, and other prohibited visual content.
• Develop text classification models to detect harassment, cyberbullying, hate speech, threats, abusive language, and sexual harassment.
• Implement audio analysis models for speech recognition, abusive language detection, and identification of prohibited audio content.
• Evaluate open-source models and commercial AI moderation APIs before deciding whether to fine-tune existing models or develop custom models.
• Build and maintain datasets for model training, validation, and testing.

2. Automated Video and Livestream Moderation

• Build systems that analyze videos using frame sampling, scene detection, and video classification.
• Develop near-real-time moderation pipelines for livestreams.
• Detect policy violations in video, audio, text, and other relevant content signals.
• Implement configurable moderation thresholds and automated enforcement rules.
• Optimize inference latency to support large numbers of concurrent video uploads and livestreams.

3. Natural Language Processing

• Develop models for detecting harassment, hate speech, threats, scams, spam, and other policy violations in comments, captions, messages, and descriptions.
• Support multilingual moderation and identify abusive content expressed through slang, misspellings, code words, and contextual language.
• Integrate speech-to-text systems to analyze spoken content in videos and livestreams.

4. AI Infrastructure and Backend Integration

• Integrate moderation models with backend microservices and APIs.
• Develop asynchronous processing pipelines using message queues and background workers.
• Containerize inference services using Docker and support scalable deployment infrastructure.
• Optimize GPU/CPU utilization, inference throughput, response time, and infrastructure costs.
• Integrate moderation workflows with video uploads, content publishing, livestreaming, reporting, and account enforcement systems.
• Design reliable APIs and event-driven services for moderation results and enforcement actions.

5. Model Evaluation and Continuous Improvement

• Measure precision, recall, false-positive rates, false-negative rates, and performance across different languages and content categories.
• Develop automated evaluation pipelines and regression tests.
• Implement feedback-driven model improvement and controlled retraining workflows.
• Monitor model drift and identify changes in content patterns.
• Maintain versioning, rollback, monitoring, and deployment processes for trained models.

6. Security, Privacy, and Compliance

• Implement safeguards for child sexual exploitation, credible threats, and other high-severity violations.
• Establish appropriate confidence thresholds and escalation rules for ambiguous or high-risk cases.
• Ensure secure handling of user-generated content and training datasets.
• Follow applicable privacy, data protection, and online safety requirements.
• Prevent unauthorized access to moderation data and model infrastructure.

Required Technical Skills

• Strong Python programming skills.
• Professional experience with machine learning and deep learning.
• Experience with PyTorch, TensorFlow, or equivalent ML frameworks.
• Computer vision and image/video classification.
• NLP and text classification.
• Experience with pretrained models, transfer learning, fine-tuning, and dataset preparation.
• Familiarity with transformer architectures and multimodal AI.
• Experience deploying models through REST APIs or inference services.
• Experience with Docker, Git, and production deployment workflows.
• Knowledge of model evaluation, data pipelines, performance optimization, and monitoring.