Computer Vision / Deep Learning Engineer

We are looking for a Computer Vision Engineer with strong experience in deep learning and applied computer vision systems. The role involves working on challenging CV problems in real-world environments. We consider Middle and Senior engineers, with scope adapted to experience. Responsibilities - Develop, train, and optimize deep learning models for: - image retrieval - image matching (keypoint detection and matching) - auxiliary perception tasks supporting the main pipeline - Evaluate models using quantitative metrics, including: - retrieval quality (mAP, Recall@K) - matching performance (precision/recall, repeatability) - end-to-end system metrics (accuracy, latency, robustness) - Optimize models for production deployment using modern toolchains: - ONNX / TensorRT / OpenVINO / edge acceleration where applicable - model compression techniques (quantization, pruning, distillation) - latency, memory, and throughput optimization - Work with large-scale visual datasets and descriptor-based representations - Collaborate with engineering teams to integrate models into production systems Required Qualifications - 2–4+ years of experience in Computer Vision / Deep Learning - Hands-on experience with keypoint detection and matching models (e.g. SuperPoint, R2D2, DISK, LightGlue, SuperGlue) - Experience with image retrieval or metric learning systems - Strong understanding of geometric and motion-related computer vision concepts: - keypoint detection and description - image matching and geometric verification (RANSAC, homography, PnP) - pose estimation and refinement techniques (PnP, bundle adjustment, pose graph optimization) - optical flow and frame-to-frame tracking - vector search / ANN methods for descriptor retrieval - Strong Python skills (PyTorch and scientific computing stack) - Ability to read and understand inference code in C++ Nice to Have - Experience with noisy or imperfect real-world datasets - Self-supervised or unsupervised learning methods (contrastive learning, homography supervision, etc.) - Experience optimizing models for edge deployment (quantization, pruning, distillation) - Familiarity with FAISS or similar vector search systems - Experience with optimization libraries for geometric problems (bundle adjustment, pose refinement) - Understanding of real-time constraints (latency, memory, CPU inference budgets on ARM or low-power x86 systems) Відгукнутись на вакансію