Inside look into how SpringML executes ML projects using TensorFlow

SpringML is passionate about helping enterprises in their data journey. We take pride in being the change agents for enterprises in their AI adoption. Building enterprise grade ML deployment is an arduous task which at times is more art than science. In this blog, I’ll walk you through how SpringML executes machine learning projects. We have been an early Google Cloud partner since 2015 and have been leveraging the Google Cloud machine learning capabilities using products such as BQML, Vision AI, Video AI and the AI platform itself to train and host ML models. Even 5 years back, executing projects that analysed text or images were considered to be out of reach by many enterprises. Today, with availability of pre-trained models APIs, AutoML and end to end AI platform makes it possible to build AI powered applications in days.


SpringML has used TensorFlow extensively on various machine learning projects. City of Memphis saw a reduction of over 20000 man hours in predictive maintenance of potholes. “Working with Google and SpringML to reduce potholes and urban blight using machine learning and artificial intelligence was an easy decision.” says Mike Rodriguez, CIO, City of Memphis

Geotab was able to accurately predict vehicle downtimes by analyzing telematics data from over 1.4 million vehicles. “SpringML helped us brainstorm how to leverage machine learning to offer greater value and insights to our customers,” says Mike Branch, Vice President of Data and Analytics for Geotab.

It’s no surprise that many of the ML platforms rely on TensorFlow as the machine learning platform under the hood. We regard TensorFlow as a mothership of all ML frameworks. Hence, we are super excited to be the launch partner for the TensorFlow AI Service Partners program to provide best in class data science engineering services to enterprise.

We’re proud to have some of the best TensorFlow certified developers in our team

Secret is in the Recipe

(Pre-trained model) APIs% + AutoML % + TensorFlow % = 100% Customer Success

Using a pre-trained model provided as an API, e.g Vision API, Video AI API is one of the fastest ways to deploy an ML project. However, as our experience has shown, most times APIs do not meet all the requirements and we have to design custom ML models in TensorFlow. Let me highlight with a recent example.

One of our customers who manage a large portfolio of image asset repositories required to classify images automatically. Given that they receive several thousand images daily, purpose built AI-powered solutions that categories images was the way to go.  Key challenge was to categorize based on the 120 attributes that customers had defined their taxonomy. Using APIs would mean that we may not be able to meet all the requirements and on the other side, building it all custom would make it a compute and resource intensive implementation.

We built a ML pipeline that brought together both best of Vision API, AutoML Vision and custom TensorFlow to not only meet all the requirements but also do it in a cost effective way.

Mobile - TensorFlow Lite, Web Applications - TensorFlow.js

TensorFlow Lite is an open source deep learning framework for on-device inference. One of our customers, a major tire manufacturer, wanted our help to read the ‘DOT’ characters on tires using a mobile app. Key requirement is to have the ML model executed in the mobile device itself. We leveraged the TensorFlow Lite to host our model on the mobile app. This solution provided a real time inference engine on the mobile device. Similarly, TensorFlow.js is a library for machine learning in JavaScript. This has allowed us to build ML models in JavaScript, and use ML directly in the browser and has been our goto choice for smart application development.

Future is an open cloud. TensorFlow is open source

We strongly believe that TensorFlow with a vibrant community support and comprehensive framework that includes TensorFlow Lite, TensorFlow.js, TFX and TensorFlow is a natural choice for everyone to develop and train machine learning models.

SpringML provides specialized services for TensorFlow developments and MLOps managed services to deploy and maintain enterprise grade ML solutions. For more information on the service offering, contact us at

About SpringML

SpringML delivers data-driven digital transformation outcomes with an experimentation and design thinking mindset. We provide consulting and implementation services and industry-specific analytics solutions that deliver high-impact business value from data. SpringML is a premier Google Cloud partner with capabilities to plan, assess, deploy, and manage data-driven engagements. We have been awarded Google Cloud specialization based on our expertise and customer portfolio for Application Development, Data Analytics, Machine Learning, and Marketing Analytics.

About TensorFlow

TensorFlow is an end-to-end open source platform for machine learning. It has a comprehensive, flexible ecosystem of tools, libraries and community resources that lets researchers push the state-of-the-art in ML and developers easily build and deploy ML powered applications. For more information, please visit this link.