Vision AI Engineer

We're looking for a Vision AI Engineer to help build nextgeneration video intelligence systems powered by modern visionlanguage models. You'll work across the full video understanding stack-combining multimodal foundation models with established analytics approaches to deliver reliable, productionready AI solutions.

Key Responsibilities

. Build endtoend video analytics pipelines using visionlanguage models.

. Finetune and adapt foundation models for domainspecific video understanding.

. Integrate VLM reasoning with traditional video analytics components.

. Develop and maintain inference pipelines for video and multimodal data.

. Deploy and optimize models for scalable, highperformance production use.

. Diagnose model issues and strengthen system stability and robustness.

. Collaborate with product and engineering teams to deliver AI-driven features.

Required Qualifications

. Strong background in computer vision, video analytics, or AI engineering.

. Practical experience with visionlanguage and videolanguage architectures.

. Hands-on experience finetuning, evaluating, and deploying deep learning models.

. Familiarity with foundation models such as CLIPbased architectures, BLIP/BLIP2, and opensource VLMs (e.g., QwenVL, InternVL).

. Proficiency in Python and deep learning frameworks (e.g., PyTorch).

. Solid understanding of CNNs,Transformers, and attention mechanisms.

. Experience with model optimization techniques (quantization, batching, memory strategies).

. Experience deploying models on Docker, cloud platforms, or onprem GPU systems.

Preferred Qualifications

. Master's or PhD in Computer Vision, Machine Learning, AI, or related fields.

. Experience with realtime or nearrealtime video analytics.

. Familiarity with traditional VA methods (detection, tracking, motion analysis).

. Exposure to MLOps workflows(versioning, CI/CD, monitoring).

. Interest in modern VLM and video understanding research.

What We Offer

. Opportunities to work on cuttingedge multimodal AI technologies.

. Ownership of productionscale video intelligence pipelines.

. A collaborative environment that blends research and engineering.