Senior Computer Vision Engineer

About the role

SoftServe is looking for a Senior Computer Vision Engineer to shape the technical vision for visual AI solutions. You will lead the transition from research to production, delivering scalable systems in image generation, multimodal intelligence, and 3D scene understanding.

Key responsibilities

• Define and communicate the technical roadmap for computer vision, generative AI, and 3D scene understanding aligned with business goals.
• Design, build, and ship production‑ready vision pipelines using PyTorch, ONNX, TensorRT, Triton and cloud platforms.
• Collaborate with Data Scientists, MLOps Engineers, ML Architects, and clients to translate complex requirements into scalable AI solutions.
• Architect diffusion models, transformers, and vision‑language models optimized for deployment.
• Establish and improve engineering processes covering model training, cloud infrastructure (AWS, Azure, GCP) and MLOps best practices.
• Mentor and grow a team of Computer Vision Engineers, sharing expertise in classical and deep‑learning techniques.
• Represent SoftServe in publications, client engagements, industry events and research communities.

Required profile

• Extensive experience delivering end‑to‑end computer vision systems in production.
• Deep knowledge of classical vision (camera calibration, feature matching, homography, geometric scene understanding).
• Advanced expertise in generative and multimodal AI (diffusion models, transformers, vision‑language models).
• Proven track record with model optimization and deployment using ONNX, TensorRT or Triton on AWS, Azure or GCP.
• Leadership experience guiding and mentoring engineering teams.
• Strong stakeholder management and ability to present technical strategies to executives.
• Academic background in Computer Science, Applied Mathematics, Physics or related field.
• Upper‑intermediate English communication skills.

Required skills

• Python
• PyTorch
• ONNX
• TensorRT
• Triton Inference Server
• AWS
• Azure
• Google Cloud Platform
• MLOps
• Diffusion models
• Transformers
• Vision‑language models
• Camera calibration
• Feature matching
• Homography estimation
• Geometric scene understanding