I help companies leverage the untapped value in their unstructured data (data such as images, videos, and reviews) by helping them identify data attributes that can drive their businesses; design processes that enable extracting and gathering the necessary data; and developing solutions for building this data, including the use of crowd workers.For businesses who want to leverage machine learning, one of the biggest challenges is developing the data you need to train an accurate ML model. Data scientists spend 80% of their time building and managing the data they use for training. The problem is even more acute when working with unstructured data that requires human annotation or labeling. If you’re training a computer vision system to detect inventory levels, you need to start by having humans label training images so that you have ground truth data that can be fed into your model. Alternatively, if you want to build a model to capture trends from customer reviews, you need to start by having someone identify the relevant attributes in those reviews. The process of gathering and managing this data requires time and experience in user interface design, quality management, and data flows that most data scientists typically don’t have. This can make it difficult to even start many ML projects.I spent over a decade at Amazon tackling challenging problems in data management including three years helping customers of Amazon Mechanical Turk and SageMaker Ground Truth find ways to leverage crowd workers to accomplish their goals. I have a breadth of experience using AWS technologies, working with public crowds like MTurk, and efficiently using private (internal) crowds. Whether you’re trying to build data to explore ML solutions or simply have untapped pools of data you’d like to take advantage of, I’m happy to help you realize your goals.