AI / Machine Learning in India



  • Does anyone in the community come across companies or groups that are doing cutting-edge work in AI or Machine Learning in India? I've come across a few companies like www.paralleldots.com and www.madstreetden.com

    Any others?

    Did you know notCRUD has cool signatures? Use them by going to your User Profile > Edit.

    Personal blog: paraschopra.com/blog/

    2


  • There's Airwoot that's done a lot of cutting-edge research into algorithms for NLP and sentiment analysis to power social CRM for big brands like Airtel and Flipkart. On the same note, I'd also call out Belong.co for using some really good custom recommendation engines to power recruitments.



  • I came across http://www.siliconbot.com/. These guys have built a text classifier to label documents or huge sets of data and then classify it based on different business segments.



  • Did you know notCRUD has cool signatures? Use them by going to your User Profile > Edit.

    Personal blog: paraschopra.com/blog/

    2


  • This post is deleted!

    ga___jit

    2


  • @khitchdee Thanks for checking us out. I agree with you about the clarity of value proposition of the product, it is specially important when the underlying technology is complex. For this product we want to closely follow the UNIX philosophy of "do one thing, do it well".



  • This post is deleted!

    ga___jit

    0


  • www.targetingmantra.com is using machine learning data to give online shoppers a personalized experience on web and mobile.



  • @parasarora said:

    www.targetingmantra.com is using machine learning data to give online shoppers a personalized experience on web and mobile.

    Have you blogged about what techniques you use? Will be interesting to read. I'm assuming you use some version of collaborative filtering?

    Did you know notCRUD has cool signatures? Use them by going to your User Profile > Edit.

    Personal blog: paraschopra.com/blog/

    0




  • We at SDSLabs, IIT Roorkee are using Machine Learning for recommending content to users on our various platforms:

    • ErdÅ‘s is using Forsit (open sourced) engine which we build ourselves to recommend problems to our users.
    • Muzi, our internal intranet music player uses Machine Learning to recommend songs to users.
    • We used ML to build a project to recognize content written on our white board by anyone in any handwriting and update that to our workflowy.
    • Other examples include tagging systems for our various applications.


  • When we were in university, we implemented nearly close to AI in Traffic control system, we tried to deploy it but authorities said not is not enough for to deploy though



  • Front end of our project is in GitHub and front end was powered by wordpress and WP REST APIs



  • @rudimk fwiw, Flipkart does not use them (contrary to what is mentioned). Airtel uses another system made by another Indian firm.



  • There are quite a number of "internal platforms" built by both Snapdeal and Flipkart for ML. The target customers are internal teams at these large orgs, but both of them organise data science group meetups with external folks regularly.



  • We at www.rightrelevance.com use Big data & machine learning for various Tasks:
    A One line brief about us: We mine through 300M twitter users, extract top 2.5M Influencers, Partition the 2.5M Influencer graph into ~40K deep interest Topic space, & use the influencer tags to bring all their articles into respective topic categories.
    We use batch processing & stream processing, with ML at various stages:

    1. Topic Modelling: Mining through approx. 30M Articles in the monthly cycle to update our Topic space.
    2. Disambiguation System: Using ML to disambiguate terms & acronyms. e.g.: if nlp is natural language processing or neuro linguistic programming.
    3. Using customized Page Rank for community based scoring & ranking.
    4. using ML to identify if a twitter handle belongs to a person or organisation.
    5. Streaming pipeline which uses ML (topic modelling) for classification of articles we process & add in our system.
    6. Locations: Using RAW data & our internal data models for identifying location affinity.

    Technologies we use: Apache Solr, Elastic Search for search, Apache Hadoop for batch Jobs, Apache Storm for stream processing, Mongo for NoSQL, Rabbit & Kestrel are our preferred queues, Redis for fast lookup, & tons of other stuff here & there.

    PS: If you have a use case for API's, we'll be happy to provide for topical content/Influencers/conversations, etc.
    A login via twitter to check analysis followers/following will help understand the accuracy of influencers & system better.


Log in to reply