Build and ship computer vision capabilities by combining classical CV, machine learning, and vision-language models, from identifying high-impact workflows to training models and monitoring production pipelines.
Computer Vision Engineer (Video Analytics, Python) - OP02259-OS
π Description
- Develop and optimize GPU-accelerated video inference pipelines for real-time processing.
- Implement, evaluate, and improve object detection models such as YOLO.
- Build temporal smoothing and tracking logic for safety event detection.
- Optimize inference with TensorRT, ONNX, CUDA, and GPU profiling tools.
- Integrate event-driven APIs, Azure Event Hub, Blob Storage, and internal services.
- Add telemetry and fail-safe mechanisms; support dataset curation, model evaluation, and containerized production deployments.
π οΈ Requirements
- 7+ years of software engineering experience with strong backend skills.
- 3+ years of experience shipping computer vision solutions to production.
- Strong Python proficiency; experience with OpenCV, PyTorch, async I/O frameworks, and API integrations.
- Hands-on experience with YOLO/Ultralytics or similar object detection frameworks.
- Experience detecting, labeling, and tracking objects in video.
- Knowledge of frame sampling, confidence thresholds, temporal filtering, and multi-camera aggregation.
- Experience optimizing GPU inference, including batching, stride, TensorRT, CUDA, quantization, and throughput tuning.
- Effective communication and collaboration with engineering and product teams.
Full job description
New opportunities in Latin America!
π’ Dev.Pro is a US-based software development company helping tech companies build their products and grow their business. Weβre remote-first, and we work with independent contractors in over 55 countries, so youβd be collaborating in a multicultural environment on some really interesting projects.
We invite a skilled AI Engineer to join our team in building GPU-accelerated video analytics for real-time safety monitoring across large fleets and industrial environments. In this role, youβll help design and optimize pipelines that detect and track actionable safety events in high-volume video streams. Youβll work on real-world systems that make an impact, collaborating with engineering and product teams to deliver reliable, high-performance solutions.
π© What's in it for you:
- Collaborate with top engineers on an innovative, technically challenging computer vision project
- Build impactful systems that enhance safety across large fleets and industrial operations
- Gain hands-on experience with GPU-accelerated video analytics, temporal tracking, and real-time safety monitoring
β Is that you?
- 7+ years of software engineering experience with strong backend skills
- 3+ years of experience shipping computer vision solutions to production
- Strong Python proficiency
- Experience with OpenCV, PyTorch, async I/O frameworks, and API integrations
- Hands-on experience with YOLO/Ultralytics or similar object detection frameworks
- Experience detecting, labeling, and tracking objects in video
- Solid knowledge of frame sampling, confidence thresholds, temporal filtering, and multi-camera aggregation
- Experience with GPU inference optimization (batching, stride, TensorRT, CUDA, quantization, throughput tuning)
- Ability to communicate and collaborate effectively with engineering and product teams
- Upper-Intermediate+ English level
Desirable:
- Experience with Azure Event Hub, Blob Storage, Application Insights, or similar
- Familiarity with Docker, cloud deployments, and production monitoring systems
- Experience in temporal/sequence analysis for event detection
- Experience with video analytics in safety, compliance, or industrial/transportation environments
π§©Key responsibilities and your contribution
In this role, youβll develop and maintain robust backend solutions that process video data and deliver actionable insights using pre-trained computer vision models.
- Develop and optimize GPU-accelerated video inference pipelines for real-time processing
- Implement, evaluate, and improve object detection models (YOLO or similar)
- Build temporal smoothing and tracking logic for safety event detection
- Optimize model performance using TensorRT, ONNX, CUDA, and GPU profiling tools
- Build and maintain integrations with event-driven APIs, Azure Event Hub, Blob Storage, and internal services
- Add metrics, logging, telemetry, and fail-safe mechanisms for resilient inference jobs
- Contribute to dataset curation, labeling, and model evaluation; fine-tuning/training as needed
- Support containerized deployments (Docker) and assist with monitoring and scaling production workload
Our next steps:
β Submit a CV in English β β Intro call with a Recruiter β β Interview β β Signing the Contract
Interested? Find out more:
By submitting application, I've read and agree to the Dev.Pro Private Policy and Consent on Processing of Personal Data. I agree to have my Personal Data being stored and processed by Dev.Pro.
Similar jobs
Search more Computer Vision Engineer jobsProductionize 3D scene-understanding systems for embodied AI, translating perception research into reliable pipelines that label environments and support downstream simulation, policy learning, and customer workflows.
Develop and improve computer vision models for real-world traffic detection and tracking, owning the full cycle from data strategy and reproducible training to field validation and edge deployment.
Own and advance localization technology by designing production-grade computer vision systems, scaling visual positioning and 3D mapping algorithms, and turning research into reliable, high-performance products.
Own end-to-end computer vision pipelines, develop and validate image-processing and machine-learning algorithms, and lead rigorous research projects for space domain awareness and space situational awareness.
