Machine Learning Engineer - Python

<p></p><p>Key Responsibilities : </p><p><br/></p><p>- Develop, train, and deploy machine learning models, focusing on performance, scalability, and maintainability.</p><p><br/></p><p>- Build and manage end-to-end ML pipelines, including data ingestion, transformation, training, evaluation, and deployment.</p><p><br/></p><p>- Work with big data processing pipelines to handle large-scale datasets for training and inference.</p><p><br/></p><p>- Generate and prepare datasets in formats such as TFRecords, LMDB, or similar formats.</p><p><br/></p><p>- Perform model evaluation, benchmark various versions on different datasets, and generate comprehensive reports.</p><p><br/></p><p>- Conduct data analysis and visualization to identify trends, outliers, and areas for improvement.</p><p><br/></p><p>- Debug issues related to data processing, model training, and Python scripts.</p><p><br/></p><p>- Utilize development tools like Git/Gerrit, static/dynamic analysis tools, code coverage, and performance profiling tools.</p><p><br/></p><p>Must-Have Skills : </p><p><br/></p><p>- Proficient in Python programming language.</p><p><br/></p><p>- Experience with TensorFlow or Keras frameworks.</p><p><br/></p><p>- Hands-on experience in developing and maintaining ML pipelines.</p><p><br/></p><p>- Strong debugging skillsespecially around Python errors, data ingestion, and compilation issues.</p><p><br/></p><p>- Good understanding of data formats used in ML model training.</p><p><br/></p><p>- Exposure to model benchmarking and performance Skills : </p><p><br/></p><p>- Experience in the Vision Domain (e.g., image classification, object detection).</p><p><br/></p><p>- Exposure to synthetic data generation for training machine learning models.</p><p><br/></p><p>- Experience building scripts for data analysis, annotation, or automated filtering of image or text data.</p><p><br/></p><p>- Familiarity with data versioning practices and tools.</p><p><br/></p><p>- Experience with tools for data annotation, AI data visualization, and labeling.</p><p><br/></p><p>- Capability to identify failure classes or problematic data patterns and report actionable insights.</p><p></p> (ref:hirist.tech)