ML Engineer (Data Engine)

Why work at Higgsfield AI? Higgsfield AI is the fastest-scaling generative AI company in history, hitting $500M in annual revenue run rate, 25M+ users worldwide, 6M+ generations per day, and powering 390 of Fortune 500 brands. We're building at the absolute frontier of AI-powered video creation and next-generation creative tools. Joining Higgsfield means becoming part of a high-impact team shaping the future of AI-native experiences, at a company that isn't just moving fast, but rewriting what fast looks like. What you’ll do - Analyze motion and physical dynamics in video, including optical flow, camera vs. object motion, temporal consistency, and physical plausibility such as dynamics, collisions, gravity, and deformation. - Train models that will be used in post-training, including reward models. - Build video classifiers for motion types, shot types, genres, and quality, using both classical approaches and VLM/embedding-based methods. - Develop methods for creating multi-shot video sequences: shot-boundary detection, assembling coherent multi-shot sequences with consistent characters and scenes, and generating captions for individual shots. - Work with audio: speech/music/sound-event detection, audio-visual synchronization evaluation (lip-sync and sound-to-action alignment), audio-quality filtering, and audio-track captioning. - Apply classical and learned data-curation techniques, including near-duplicate detection, heuristic filtering, VLM-based captioning, synthetic data generation, and dataset versioning. - Build and run large-scale distributed processing pipelines. Requirements - Strong PyTorch skills and hands-on model training experience, including fine-tuning and deploying classifiers, embedding models, and VLMs. - Deep knowledge of video-processing methods, including motion analysis (optical flow, tracking, camera-motion estimation), shot-boundary detection, and temporal consistency and quality evaluation. - Experience building and validating classifiers