Realtime 3D Vision ML Engineer — Edge AI & Confidence Calibration
Summary
Tyndall FX is hiring an ML engineer in Paris to build real-time 3D/video vision models for field deployment: fine-tuning large pretrained models with parameter-efficient methods (LoRA), generating synthetic training data from 3D reconstructions, and compressing models for edge hardware via quantisation and ONNX/TensorRT, while owning evaluation, calibration, and CI/CD-tested delivery.
Tyndall FX is seeking a Machine Learning Engineer to develop AI models trained on 3D and video data. You will adapt large pretrained models for real-time field use, generate training data, and ensure results carry trustworthy confidence levels.
You will collaborate with ML, CV, 3D graphics and mobile teams to push production-ready solutions, emphasizing robustness, tests, and CI/CD practices.
Fine-tune large pretrained vision/video models on 3D data with parameter-efficient methods like LoRA Design and train models for 3D and video understanding Own evaluation and calibration ensuring trustworthy results in the field Distill and compress models to real-time hardware using quantisation and ONNX/TensorRT Generate synthetic data from 3D reconstructions and close the data gap Build training/evaluation pipelines, including on-premise pipelines Take prototypes through to product: spec, integrate, release Ensure product quality via tests and CI/CD Master's or PhD in Computer Science, ML, or related field 4+ years in ML/vision, incl. 2+ years in industry Strong Python and good C++ skills Experience fine-tuning large pretrained models and optimizing for speed Shipping software products with CI, tests and customer focus English proficiency; French is a plus Bonuses and Stock Options Hybrid work flexibilitySkills
Sign in to read and join the discussion.