Machine Learning Engineer — Computer Vision(Optometry Domain, Mobile Focus)

Job Description

Lenskart is seeking a Machine Learning Engineer specializing in Computer Vision. You will lead the design, fine-tuning, optimization, and deployment of computer vision models with a focus on OCR (Optical Character Recognition), and desktop application automation for the optometry space. The ideal candidate will transform prescription forms, patient documentation, and clinical workflows into automated AI-driven processes that enable efficient optometry eye tests .

You will collaborate with product managers, optometry experts, and software engineers to bring cutting-edge computer vision and automation solutions into real-world clinical workflows, enabling seamless integration with desktop applications and patient engagement workflows.

Must Haves

• 5+ years of overall experience with 2+ years of experience with vision models.
• Proven experience with deep learning and computer vision frameworks (e.g., OpenCV, PyTorch, TensorFlow, OpenCV, Yolo, or similar vision libraries).
• Strong Python programming skills, especially in image processing (e.g., OpenCV, PIL/Pillow), data manipulation (Pandas, NumPy), and automation scripting for desktop applications (e.g., PyAutoGUI, pywinauto).
• Experience fine-tuning and deploying OCR models (e.g., Tesseract, EasyOCR, PaddleOCR) and document understanding models (e.g., LayoutLM, Donut) for extracting structured data from medical forms and prescriptions.
• Prior exposure to model evaluation using metrics such as character accuracy, word accuracy, precision/recall for OCR, and ability to benchmark performance on domain-specific datasets.
• Track record of deploying computer vision models in production environments, including desktop application integration, and experience with model optimization techniques for inference speed.
• Experience building data pipelines for vision tasks, including data augmentation, annotation workflows, and periodic model retraining with new labeled data.
• Ability to implement visualization and debugging tools for computer vision models (e.g., GradCAM, attention maps, bounding box visualizations) and communicate results to non-technical stakeholders.

Should Haves

• Experience building RESTful APIs for serving computer vision models and integrating with desktop automation frameworks (e.g., Windows COM, AppleScript).
• Exposure to basic devops practices (CI/CD pipelines for ML, model version control).