Must have

• Computer vision

• Video processing

• Python

• C++

• PyTorch

• TensorFlow

• SDK

• English

Nice to have

• Kafka

• Git

• GitHub

• GitLab

• Kubernetes

Requirements description

• Proven experience across the full ML/CV pipeline - from design and training to production deployment
• Strong algorithmic thinking and solid implementation skills.
• Hands-on experience with computer vision/AI systems and real-time image/video processing.
• Experience in model optimization and deployment in production environments.
• Proficiency in Python and C++ programming.
• Experience with frameworks and libraries such as PyTorch, TensorFlow, TensorRT.
• Familiarity with streaming and deployment technologies (RTSP, NVIDIA DeepStream SDK, Kafka).
• Experience with Git-based version control (GitHub/GitLab).
• (Nice to have) Knowledge of Kubernetes or containerization tools.

Offer description

Devire IT Outsourcing is a form of cooperation dedicated to IT professionals, based on the principles of own business - B2B, implementing projects for clients running innovative and modern projects.

With our Client - a company specializing in scalable backend development, Big Data, cloud services, data engineering, and DevOps/SRE consulting, offering end-to-end software lifecycle support and agile project management for efficient outsourcing - we are looking for Computer Vision / AI Expert.

The project focuses on advanced AI-powered security and analytics solutions. One example is an anti-theft system capable of tracking a specific person from the moment they enter the store, integrating visual data from multiple cameras, and enabling person re-identification across different zones. Another key area is inventory monitoring - recognizing the quantity of products on shelves through image analysis.

Core Challenges

• Real-time multi-camera person detection, identification, and anonymous tracking

• Body pose detection and action classification

• Efficient publishing/consuming with Kafka

• Person re-identification across camera zones

• Low-latency and real-time processing systems

• Additional benefits package (medical healthcare, sports membership, etc.)

• Long-term cooperation

• Flexible working hours

Your responsibilities

1. Contribute end-to-end - from design and prototyping to deployment in production environments (not limited to Jupyter notebooks).
2. Apply strong algorithmic skills to develop efficient, scalable solutions.
3. Design, train, and optimize deep learning models (primarily CNNs) for real-time image and video analysis.
4. Integrate models with streaming frameworks such as NVIDIA DeepStream SDK, RTSP, and Kafka.
5. Work with inference acceleration tools (e.g., TensorRT) and deploy AI models.
6. Collaborating with a team of engineers and DevOps developers to implement solutions in production environments (on the edge/on the cloud).