Da Vinci

Da Vinci is Xavient Digital’s Internal AI Platform, which is close to my heart.

It powers our Big Data, BI and analytics capabilities and its offerings are more tuned towards creation and support of algorithmic business and its needs; and as we know todays’ algorithmic businesses are propelled by two distinct trends.

The first significant shift that we are noticing today is that our clients continue to accumulate massive piles of data anywhere around a couple hundreds of Terra bytes to start with. And the other compelling trend that we observe is majority of the data that drives today’s digital business, is NOT coming from their data centers, I mean all these transformative data, rather new kind of data is coming from their cloud, from their mobile, their social networks and through connected devices and sensors. Typically this new data sits outside the perimeters of their data centers.

So engaging with this new type of data and capitalizing on this physical digital convergence is what we specialize on with Da Vinci, the Cognitive analytics practice. This is where we see lot of traction.

So I have started experimenting and piloting this ‘humans Vs machines’ strategy through Da Vinci.

We do a whole bunch of descriptive and diagnostic analytics, predictive and prescriptive analytics, undertake all four types of engagements and deal with the whole spectrum, all the tactical 12 forms of data right from transactions, logs, geo spatial, social, web, chat, email, sensor logs, free form text, voice, video and images we are building accelerators for all these 12 areas.

With Da Vinci, we continue to build extreme expertise to pick and choose classification, clustering models, neural nets and recommendation engines as appropriate to enhance customer experience. And do targeted marketing, increase process efficiency through Robots, (RPA robots), cost reduction, customer churn, risk isolation and so on and so forth.

And not only that, as we get deeper into Predictive business, Da Vinci’s main motive, main objective is to help our customers to be ahead of the crowd, to be ahead of the curve and predict a place, create a capability where the rest of the world would arrive eventually, and would catch up eventually.

Following is a subset of reusable models Da Vinci has built:

• CNN based Voice to Text model
• CNN Based intent classification
• MultiGPU - RNN Model
• MultiGPU - Character Iteration
• LSTM - Generative - Generates text, one character at a time.(eg: Shakespeare)
• Video Generator
• Video Classification
• CNN Image Recognition 
• RNN based Speech Recognition
• RNN -asynchronous data iteration 
• RNN News Iterator
• RNN Seq2Seq predictor/Addition
• RNN classify univariate time series 
• Transfer learning model
• Un Supervised - DBN - Deep Belief Network
• Remote UI
• Spark EMR Model
• Spark MLP Model
• Spark LSTM
• DataVec model - pre processing operations on some simple CSV data.
• DataVec Transform - pre processing/aggregation operation