D

dreamerlabs

AI-driven interactive entertainment with generated content

Senior AI Engineer - Trust and Safety

Applied Machine LearningConfirmed live in the last 24 hours
Full-timeSydney NSW, AustraliaA$250k - A$400k /yr

About the Job

Requirements

  • Demonstrated ability to build a technically complex product at real scale or solve a difficult problem in depth, and clearly explain the work and its challenges.
  • Product judgment, including talking to users, identifying worthwhile work, and supporting decisions with evidence; customer-facing product ownership is preferred.
  • Ability to take end-to-end ownership of an ambiguous problem, ship a product without it being decomposed into assigned tasks, and own post-launch outcomes.
  • Trust and safety experience, ideally building moderation or safety models at scale and handling sensitive data properly.
  • Ability to evaluate impact and cost, run experiments, assess results, and balance rigorous implementation with simpler solutions when appropriate.
  • Ability to explain thinking succinctly using data and challenge ideas constructively.
  • Curiosity about AI models and tools and ability to assess how they could change the product.
  • Motivation to achieve high standards and take responsibility for important work.

Responsibilities

  • Define the content and behavior the product's safety models need to detect.
  • Work with labellers to prepare data for safety models.
  • Train and evaluate safety models on the company's GPUs, deploy them to production, measure their performance, and improve them.
  • Own model accuracy, latency, and cost across high-volume image and message workloads.
  • Improve detection of prohibited content in user prompts and generated images while reducing false positives without increasing missed content.
  • Use large language models and classifiers to detect prohibited content in user messages and AI responses.
  • Reduce the moderation-model retraining, evaluation, shadowing, and deployment cycle to a week or less.
  • Automate parts of content review and user banning with appropriate safeguards.

Desired Qualifications

  • Vision classifiers or vision-language models tuned for precision at very high recall on prohibited image content.
  • Distilling large language model labellers into small fine-tuned models for policy detection.
  • Using large-language-model-assisted labelling and active learning to ensure human reviewers focus on difficult cases.
  • Retraining classifiers on new abuse patterns within days using synthetic data and evaluations on real traffic.
  • Detecting users who gradually steer conversations toward prohibited content.

About the Company

dreamerlabs

AI-driven interactive entertainment with generated content

Dreamerlabs creates consumer AI entertainment products focused on interactive storytelling. Its platform lets readers influence a story by editing a highlighted phrase, which triggers regenerated video, images, and narration to match the new wording. The product works through an internal production stack that handles fine-tuning, story worlds, characters, audio playback, and media generation to deliver updated scenes as users edit the text. Dreamerlabs distinguishes itself by offering a built-in production pipeline and the ability to regenerate media in real time from user edits, supported by ongoing hiring and visible open roles. The company's goal is to provide engaging, customizable AI-driven entertainment experiences that let users shape stories and characters while expanding its library of story worlds and audio-visual content.

Company size
11-50
Company stage
Not listed
Total funding
Not listed
Headquarters
Sydney, Australia
Founded
2025

Recently Posted Jobs

Explore more opportunities at dreamerlabs

See more jobs