Sparx: Open Source Machine Learning API Server

Bastin Robin (~BastinRobin)


Sparx Microservice

Sparx has been open sourced by CleverInsight Inc, Sparx helps the data scientists to deploy their ML, DL, and NLP models faster as microservice. Sparx has been built for people who don't want to touch more python code, it's a server which has been built with YAML script configuration. Sparx comes with a CLI (Command Line Utility) which can simplify the task of creating an application which runs on tornado with configurable YAML script for exposing endpoint API's. Each API's is considered to service in Sparx ecosystem.

Sparx has been used for the majority of our deployments which have really helped us to build interesting ML/DL & NLP use cases.

Slides : Sparx.ppt

Demo: Live Action


In the current scenario, we are mainly more focused on building predictive models, most of the data scientist or data analyst use Jupyter Notebook for their experimentation. Once they have successfully built the model the most important phase is to make it live via API or Microservices. Majority of analysts requires a data engineer or software engineer's support for making their model live for usages. Sparx solves that problem by enabling the analyst to initiate a high-speed API server with few command line arguments. Commands create RESTful scaffolding with the ML/DL model binding with POST requests. My talk will be focused on how we used Sparx to build our Machine Learning API services. Why we came with Sparx and Inviting contributors to this framework to make it open for python and data-science community.

Basic outline of talk

  1. Introduction to Sparx [2-4 minutes]
  2. Feature of Sparx [4-5 minutes]
  3. How easy to production ML models in Sparx - Demo [5-10 minutes]
  4. How to use Sparx via Sparx-CLI [2-4 minutes]
  5. Open Source Contribution Request [2-4 minutes]
  6. Roadmap of Sparx [4-5 minutes]
  7. Q/A - [2 minutes]

Who is this talk for?

Python developers who deal with data analysis Data analyst, R programmers developers who'd like to deploy their ML/DL/NLP models.

Key takeaways

  1. Framework build for easy deployment of Data Science Projects
  2. Learn to deploy ML models as RESTFul or Microservice in few minutes.
  3. Contribute to Sparx-Core


  • Python 2.7 and above
  • Machine Learning
  • YAML
  • API development
  • Deployment
  • CLI

Content URLs:

Speaker Info:

Hi, I am Bastin Robins .J, seasoned Data Practitioner who has successfully built data products for largest FMCG's and Retail, Banking, Telecom, Social, Government from scratch and deployed it successfully with great feedbacks. Experienced in developing products which are a combination of Machine Learning, Deep Learning, Neural networks, Collective Intelligence & Web Intelligence. Currently working as Chief Data Scientist at CleverInsight.


Id: 1213
Section: Data Science, Machine Learning and AI
Type: Talks
Target Audience: Intermediate
Last Updated: