{"id":1544,"date":"2019-02-28T03:00:10","date_gmt":"2019-02-28T03:00:10","guid":{"rendered":"http:\/\/kusuaks7\/?p=1149"},"modified":"2023-06-29T12:06:10","modified_gmt":"2023-06-29T12:06:10","slug":"four-most-important-success-factors-in-any-machine-learning-project","status":"publish","type":"post","link":"https:\/\/www.experfy.com\/blog\/ai-ml\/four-most-important-success-factors-in-any-machine-learning-project\/","title":{"rendered":"Four Most Important Success Factors in any Machine Learning Project"},"content":{"rendered":"

If you are a product manager<\/strong> and want to build something with machine learning, here’s a list of the 4 most important things to keep in mind:<\/p>\n

1. Prioritise engineering over data science<\/h2>\n

<\/p>\n

 <\/p>\n

A machine learning project is first and foremost a software project. Many data scientists have little experience building well architected, reliable and easy to deploy software<\/strong>. When you build a production system, this will become a problem.<\/p>\n

As a rule of thumb, engineers can pick up data science skills faster than data scientists can pick up engineering experience.<\/strong> If in doubt, work with the python engineer with 5+ years experience and a passion for AI, rather than the PhD in data science who is having their first go at building business applications.<\/p>\n

2. Go lean<\/h2>\n

It’s important to reduce risks early. Structure your project with concrete milestones:<\/p>\n

    \n
  1. Finished Prototype<\/strong>: Find out whether your idea is promising 1 day — 2 weeks<\/em><\/li>\n
  2. Offline tested system<\/strong>: Tune the model and rigorously test it on existing data 2–4 weeks<\/em><\/li>\n
  3. Online tested system<\/strong>: Finalise the model and test it live 2–4 weeks<\/em><\/li>\n
  4. Going live:<\/strong> Automate data updates, model training and code deployment:2–4 weeks<\/em><\/li>\n
  5. Continuous improvement: <\/strong>(optional) <\/strong>12 months<\/em><\/li>\n<\/ol>\n

    Total timeline: 1–3 months<\/strong><\/p>\n

    An experienced team should be able to follow these timelines for almost any project. Focus the team on setting up a live system in 1–3 months. After it’s live, then decide whether further improvements are worth it.<\/p>\n

    These temptations can prolong your project unnecessarily:<\/h3>\n
      \n
    • Waiting for the perfect data<\/li>\n
    • Using the wrong tools (too complex or too slow)<\/li>\n
    • Overengineering for scalability<\/li>\n
    • Endlessly playing with the algorithms (see next point)<\/li>\n<\/ul>\n

      3. The algorithm doesn’t matter<\/h2>\n

      <\/p>\n

       <\/p>\n

      Machine learning systems have lots of fascinating knobs you can play with. Don’t.<\/p>\n

      The improvements that are worth spending time on (in order of importance):<\/p>\n

        \n
      1. Get more (relevant) input data<\/li>\n
      2. Preprocess the data in a better way<\/li>\n
      3. Choose the right algorithm and tune it correctly.<\/li>\n<\/ol>\n

        The algorithm is the least important factor<\/strong>. Simply choose an algorithm that works. Endlessly upgrading the algorithm is tempting, but it will probably not give you the results you expect.<\/p>\n

        4. Communicate, communicate, communicate<\/h2>\n

        <\/p>\n

         <\/p>\n

        Share as much of the business context as possible:<\/h3>\n

        Once the engineering team starts building, they have to make a lot of choices. The better they know your priorities, the more they can make the right decisions. You should at least tell them about:<\/p>\n

          \n
        • Strategic priorities<\/strong><\/li>\n<\/ul>\n

          Is this fixing a critical issue? Will it need to work for millions of requests a day? Or is it research for a future product?<\/p>\n

            \n
          • Problems with the current process<\/strong><\/li>\n<\/ul>\n

            Does the current process take too long? Is it too inaccurate? Or is there a lot of data that simply can’t be taken into account without machine learning?<\/p>\n

              \n
            • Inputs and outputs<\/strong><\/li>\n<\/ul>\n

              Inputs: What data would you (as a human) use to make the right decisions?<\/p>\n

              Outputs: Who will consume the output? How frequently? Does it need to be real time?<\/p>\n

                \n
              • Performance metrics<\/strong><\/li>\n<\/ul>\n

                What are the most important metrics: Click through rate? Sales? ROI? False positive rate?<\/p>\n

                  \n
                • Expected accuracy<\/strong><\/li>\n<\/ul>\n

                  If you want to optimise conversion rates, then it might not be worth another 2 weeks of tuning to get 2% more accuracy.<\/p>\n

                  If you build medical diagnostic systems, then false negatives of even 1% can be unacceptable.<\/p>\n

                  TL;DR<\/h2>\n
                    \n
                  • Prioritize engineering over data science.<\/li>\n
                  • Reduce risks by going lean.<\/li>\n
                  • Don’t get distracted by the algorithm.<\/li>\n
                  • Share all business requirements with your developers.<\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"

                    A machine learning project is first and foremost a software project. Many data scientists have little experience building well architected, reliable and easy to deploy software. When you build a production system, this will become a problem. As a rule of thumb, engineers can pick up data science skills faster than data scientists can pick up engineering experience. If in doubt, work with the python engineer with 5+ years experience and a passion for AI. If you are a product manager and want to build something with machine learning, here’s a list of the 4 most important things to keep in mind.<\/p>\n","protected":false},"author":314,"featured_media":3995,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"content-type":"","footnotes":""},"categories":[183],"tags":[92],"ppma_author":[2069],"class_list":["post-1544","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-ml","tag-machine-learning"],"authors":[{"term_id":2069,"user_id":314,"is_guest":0,"slug":"markus-schmitt","display_name":"Markus Schmitt","avatar_url":"https:\/\/secure.gravatar.com\/avatar\/?s=96&d=mm&r=g","user_url":"","last_name":"Schmitt","first_name":"Markus","job_title":"","description":"Markus Schmitt is the founder and head of data science at Data Revenue, a Machine Learning Agency based in Berlin, Germany, where he builds custom end-to-end machine learning systems for Medical, Finance and Marketing clients. Before Data Revenue he developed new ventures for the company builder Team Europe and studied Mathematics & Economics at Warwick."}],"_links":{"self":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts\/1544","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/users\/314"}],"replies":[{"embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/comments?post=1544"}],"version-history":[{"count":2,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts\/1544\/revisions"}],"predecessor-version":[{"id":28966,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts\/1544\/revisions\/28966"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/media\/3995"}],"wp:attachment":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/media?parent=1544"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/categories?post=1544"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/tags?post=1544"},{"taxonomy":"author","embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/ppma_author?post=1544"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}