Sachin Rajan

As a Computer Vision Engineer at Klevu, I develop and deploy innovative solutions for image search and object detection in various domains, such as furniture, and apparel. I use cutting-edge technologies and frameworks, to create high-performance and scalable modules that improve the user experience and satisfaction of our clients.

Computer Vision Engineer at Klevu Oy (03/202 - Present, Remote):

• Developed an Image Search engine, enabling the retrieval of similar images from a set based on a query image.
• Implemented a color identification module with 90% accuracy and integrated it into the Image Search engine, enhancing performance.
• Trained a classifier for furniture object identification with close to 90% accuracy.
• Created endpoints for diverse image processing and deep learning modules using Flask.
• Trained EfficientDet-based object detection models in furniture and fashion domains.
• Integrated detection models to optimize the search speed of the Image Search engine by reducing the search space.

Computer Vision Engineer at Camcom (01/2018 - 02/2021, Bangalore):

• Developed a model with 92% accuracy for detecting helmets in a video stream to enhance rider safety.
• Designed an automated quality check system for apparel, achieving an accuracy of 90%.
• Deployed the apparel quality check solution in all warehouses of the Myntra e-commerce platform.
• Identified damages on vehicles with 90% accuracy for estimating insurance claims.
• Conducted a POC for reading registration certificates for all Indian states as part of the damage detection use case.
• Created an Annotation validator tool for various models to streamline data validation processes.
• Managed a team of computer vision engineers, including task allocation and progress monitoring.

Computer Vision Engineer at Profiledskin (07/2017 - 12/2017, Ahmedabad):

• Developed a system to measure the length of human feet from images, providing accurate footwear size suggestions.
• Obtained various measurements of shirts for suggesting users shirts with similar dimensions.
• Achieved 90% accuracy in detecting skin pixels from images.

M.Tech in Remote Sensing from Amrita School of Engineering, with a focus on acquiring expertise in key domains such as linear algebra, probability, image processing, and the fundamentals of AI. Throughout the program, I honed my programming skills in Python to implement and execute various projects. The curriculum equipped me with a strong foundation in mathematical concepts, enabling the effective application of algorithms in the field of Computer Vision.