Computer Vision Engineer

Огляд

Employeer is a defense tech company specializing in the development of innovative solutions in Embedded Systems and radio frequency (RF) engineering.

Обов'язки

• development, training, and optimization of deep learning models for image retrieval, image matching, keypoint detection and matching, and auxiliary perception tasks
• evaluation of models using quantitative metrics, including retrieval quality (mAP, Recall@K), matching performance (precision/recall, repeatability), and end-to-end system metrics (accuracy, latency, robustness)
• optimization of models for production deployment using ONNX, TensorRT, OpenVINO, and edge acceleration where applicable
• application of model compression techniques, including quantization, pruning, and distillation
• optimization of latency, memory usage, and throughput
• work with large-scale visual datasets and descriptor-based representations
• collaboration with engineering teams to integrate models into production systems

Вимоги

• 2–4+ years of experience in Computer Vision / Deep Learning
• hands-on experience with keypoint detection and matching models, such as SuperPoint, R2D2, DISK, LightGlue, or SuperGlue
• experience with image retrieval or metric learning systems
• strong understanding of keypoint detection and description
• strong understanding of image matching and geometric verification, including RANSAC, homography, and PnP
• knowledge of pose estimation and refinement techniques, including PnP, bundle adjustment, and pose graph optimization
• understanding of optical flow and frame-to-frame tracking
• understanding of vector search / ANN methods for descriptor retrieval
• strong Python skills, including PyTorch and the scientific computing stack
• ability to read and understand inference code in C++

Nice to Have

• experience with noisy or imperfect real-world datasets
• knowledge of self-supervised or unsupervised learning methods, including contrastive learning and homography supervision
• experience optimizing models for edge deployment using quantization, pruning, or distillation
• familiarity with FAISS or similar vector search systems
• experience with optimization libraries for geometric problems, including bundle adjustment and pose refinement
• understanding of real-time constraints, including latency, memory, and CPU inference budgets on ARM or low-power x86 systems
• familiarity with SLAM or VIO concepts

Умови

• full-time employment
• work from the office in Kyiv or fully remote
• official employment
• reservation from mobilization
• 24 calendar days of vacation and paid sick leave
• work in a dynamic, innovative, and large-scale team developing new products and improving existing ones