Machine Learning Engineer – Computer Vision (Bengaluru)

Machine Learning Engineer – Multimodal AI & Agentic systems

Location: Bengaluru, India

Experience: 3–6 Years (Relevant Industry Experience)

About iMerit

iMerit is a global leader in AI data solutions, enabling organizations to build reliable, production-grade Artificial Intelligence systems through high-quality data, advanced tooling, and human expertise. We partner with leading enterprises in autonomous mobility, healthcare, geospatial intelligence, robotics, and generative AI to solve complex machine learning challenges.

With over 5,500 professionals worldwide, iMerit combines cutting-edge technology with human intelligence to accelerate the development and deployment of AI systems while maintaining the highest standards of quality, security, and social impact.

About the Role

We are looking for a passionate Machine Learning Engineer – Multimodal AI & Agentic systems to join our AI Engineering team. The role combines applied research, machine learning engineering, and platform development to build production-grade systems that understand, process, and reason across multiple data modalities, including images, video, text, documents, point clouds, and sensor data.

You will develop multimodal machine learning solutions and agentic systems for data curation, intelligent annotation, model evaluation, quality validation, synthetic data generation, and automated AI workflows. The role involves applying computer vision models, Large Language Models, Vision Language Models, multimodal foundation models, and tool-using agents to solve complex real-world problems.

The ideal candidate enjoys conducting structured experiments, translating research into reliable engineering systems, and building AI workflows with appropriate evaluation, observability, safeguards, and human oversight

Key Responsibilities

Computer Vision & Machine Learning

• Design, develop, train, and optimize state-of-the-art computer vision models for tasks including object detection, semantic and instanc