Want to know if this job is worth applying to?
Jaroměř, Česko
Remote only
This position is listed on behalf of a partner company, who manages all applications and next steps. Our partner is looking for a Staff Deep Learning Engineer, State Estimation based in the United States.
This is a remote engineering opportunity focused on advancing autonomous systems that can understand their motion and localize themselves when GPS is unavailable or unreliable.
You will work at the intersection of deep learning, 3D computer vision, geometric estimation, and vision-based navigation.
The role gives you ownership of the machine learning lifecycle, from data strategy and annotation through model development, evaluation, integration, and deployment.
You will develop learned capabilities for feature matching, depth estimation, relative pose estimation, and image-to-map localization.
Your work will directly support robust localization across challenging environmental and sensor conditions.
You will collaborate closely with state estimation, software, systems, deployment, and flight-test engineers to turn research concepts into reliable production capabilities.
The position is well suited to an experienced deep learning engineer who enjoys solving complex perception and autonomy problems and translating research into deployable software.
Develop and evaluate deep learning models for feature detection and matching, visual correspondence, depth estimation, relative pose estimation, image-to-map localization, and related perception tasks.
Combine learned visual representations with geometric estimation techniques to improve localization accuracy, robustness, and recovery in challenging conditions.
Own data preparation and supervision strategies, including dataset curation, annotation requirements, labeling workflows, automated quality checks, and coverage analysis.
Select and integrate appropriate deep learning tools while building reproducible training workflows with configuration management, experiment tracking, and dataset and model versioning.
Design comprehensive evaluations covering both model-level performance and downstream localization outcomes across variations in lighting, viewpoint, altitude, terrain, weather, and sensor characteristics.
Investigate model failures and use controlled experiments to identify and prioritize improvements to datasets, supervision strategies, architectures, and system integration.
Partner with state estimation engineers to integrate learned measurements and confidence estimates into visual-inertial odometry (VIO) and terrain-relative navigation systems.
Profile models against onboard compute, memory, and latency constraints and collaborate with deployment engineers on optimization and runtime validation.
Deliver tested, documented software components and interfaces for an autonomy SDK while collaborating with software, systems, and flight-test teams.
Communicate technical assumptions, experimental findings, limitations, and design tradeoffs clearly to research and engineering stakeholders.
Help translate advanced computer vision and deep learning research into maintainable, tested, and deployment-ready capabilities.
Master’s degree in Aerospace Engineering, Electrical Engineering, Robotics, Computer Science, or a related technical field, with at least 4 years of relevant professional experience, or a Ph.D. with at least 2 years of relevant experience.
Hands-on experience designing, training, debugging, and evaluating deep learning models using PyTorch or an equivalent framework.
Strong understanding of architecture selection, loss-function design, optimization, and data augmentation techniques that preserve geometric consistency.
Solid foundations in camera models, coordinate transformations, projective geometry, and multi-view geometry.
Practical experience in one or more relevant areas such as vision-based navigation, visual geolocation, Structure from Motion (SfM), SLAM, 3D reconstruction, depth estimation, or related computer vision fields.
Strong Python programming skills and experience developing maintainable, reusable software.
Demonstrated ability to take a computer vision capability from problem definition and raw sensor data through training, evaluation, and integration readiness.
Experience building sensor-data pipelines covering ingestion, cleaning, filtering, deduplication, and dataset versioning.
Experience creating reproducible machine learning workflows involving configuration management, experiment tracking, checkpointing, and GPU performance troubleshooting.
Ability to design meaningful benchmarks, prevent data leakage across related sequences or locations, evaluate performance across operating conditions, and connect model metrics to downstream geometric or localization performance.
Experience profiling inference latency and memory consumption and assessing accuracy-versus-compute tradeoffs.
Ability to document model interfaces and preprocessing requirements and advise deployment teams on export, precision, and runtime optimization.
Strong communication skills, with the ability to explain technical findings, assumptions, and tradeoffs clearly and translate research into practical engineering solutions.
Preferred experience includes aerial imagery, geospatial data, elevation maps, or matching observations across different viewpoints, lighting conditions, seasons, or sensor modalities.
Experience with model export, quantization, TensorRT, ONNX, or embedded compute platforms is advantageous.
Experience validating perception or robotics systems on physical platforms is a plus.
Relevant publications, open-source contributions, or demonstrated delivery of production computer vision systems are valued.
Familiarity with feature correlation, cost volumes, matching techniques, stereo, optical flow, localization, or related correspondence methods is beneficial.
Experience applying learned priors to scene geometry, depth, motion, or appearance is advantageous.
Exposure to diffusion models or flow matching for computer vision, geometric inference, or conditional generation is a plus.
Aerospace and/or defense industry experience is preferred.
Strong deep learning and 3D vision foundations are the core requirement; candidates do not need to meet every preferred qualification.
Annual salary range of $200,000–$300,000, with compensation influenced by experience, skills, certifications, and work location.
Performance bonus opportunity.
Equity package for eligible full-time employees.
Comprehensive employee benefits package for eligible full-time employees.
Remote work environment within the United States.
Opportunity to work on advanced deep learning, computer vision, autonomy, and state-estimation challenges.
Exposure to multidisciplinary collaboration across machine learning, robotics, software, systems, deployment, and flight-test engineering.
Opportunity to contribute to production-grade autonomous systems and transform advanced research into deployed capabilities.
Temporary employees may receive a temporary benefits package after 60 days, subject to applicable eligibility requirements.
Offers are contingent upon a cleared background and, where applicable, reference checks.
Equal employment opportunity and reasonable accommodation support.
