Saturday, May 18, 2013

Launching "The Photon Effect"

Today I want to share one more investment under preparation with all of you - A story blog powered by photographs. Photography has been an enjoyable hobby for me, and I am now taking it to the next level. 

"The Photo Effect" is an attempt to tell stories. These stories are part of me when I travel, meet people and share experiences with the world. 

"The Photon Effect" is currently under preparation, and is going live the beginning of June 2013. 

Till then, keep reading !

Sunday, May 12, 2013

Deep Learning: The next paradigm for Machine Learning

2 posts that made me excited about Deep Learning, and how it is evolving the state of art in Machine Learning. 

For newbies, Deep Learning is a set of algorithms that emphasises on layered approach to learning. Each layer is a neural network corresponds to a concept, going all the way from high-order to a low-order. A high order concept is composed of low-order concept and multiple low-order concepts give way to a high-order concept. A concept could be a part of a feature - like learning to understand a language.

Post 1:- 

Andrew Ng, the celebrated Neural Network Researcher and the prestigious AI professor talks about his love for AI and Neural Networks. In The One Algorithm Brain - Andrew Ng discovers how the concept of "One Algorithm" could define how our brains perform complicated processing. And One day, we might finally built something near to it.

Post 2:- 

Rick Rashid, the Chief Research officer of Microsoft Research, wows audience with real-time translation of spoken words. The technology deep within this work is Deep Learning. 

The behind the scenes action of Post 1 and Post 2 lies in application of Deep Learning. Not only, it is creating a paradigm shift in creation of intelligent software, it also shows that key to such creation is learning how our own brain works. 

I am writing a blog post for allowing a layman understand about Deep Learning. Watch out for this in the next blog post.

Saturday, April 20, 2013

The world through the Glass

Google Glass is a phenomenal attempt to create a device that blurs the interface between a human and a machine. I believe this is a first step towards the complete integration, and a scary one too. The Glass as advertised by Google is an important step in the evolution of the human species, where the humans start accepting augmentation with machines as a natural progression. However, it brings out an important point on how Google as a company wants to invade in the private and personal space of each user. Google is primarily an Advertisement based company. It survives and thrives on people, you and me, using its powerful search engine, and other plethora of tools to play with information. This includes searching, recording, authoring, viewing and sharing. An important part of this engagement between the human and Google is the need for a consumption device. This engagement allows Google to profile the user, provide targeted advertisements and information. But a key problem with the engagement is that a user needs to be in interaction with the interface. A device fills the gap except places where the user does not intent to involve it. This may include the free time that we all have including interacting with our families, pursuing our hobbies, sharing love and experience for the first time etc. All this represent the pure human to human interaction. A machine in the form of a mobile phone, laptop, tablet is seen as an intrusion in these type of interactions, until Google Glass arrived.
The Glass opens up tremendous opportunity for Google to perceive the behaviour of each user in far detail than possible now.  This means, an unprecedented access to more of our human - to - human time. When the interface becomes invisible as in the Glass, there is very little a user can do to decide on the appropriateness of the content that gets communicated and shared. The interface is becoming a part of our real self. I will not be surprised if when using the Glass and playing a game of tennis, my Glass realises that I am poor in the sport, and superimpose training videos on my retina followed with advertisements of training institutes.
This is scary thought, but one whose time has come. I am all excited and scared at the same time :)

PLEASE NOTE:- The Image used here is a part of the published article "". The blog owner does not own any rights for the image.

Sunday, February 24, 2013

Why I can't replace paper ?

I read a lot. I try to find every ounce of time that I can squeeze from my schedule to read. Reading definitely does not mean, carrying a book everywhere, but finding every opportunity to go over a written piece of online and offline text. This means a greater challenge for me to find out the items worth spending time reading on. So I necessarily divide a part of my day into finding things to read and then going over it as I get time. 

Ever since I bought the Amazon Kindle, the graphite version of the stellar product back in mid-2010, its been a part of some of my reading time if not all. 6 months into Kindle, and I was still finding myself going over a shiny new printed paper. Find me in a bookstore, and I am hypnotized to get my hands on to the written text. Reading a PDF in the Kindle was always a pain. I still email most of the PDFs I wanted to read in my free time to my free kindle email address. Kindle has been a quite member of my reading journey since last 3 years. And I still see myself buying a lot of printed books, with some contribution of the easy-to-order online stores promising me 1 - 2 day delivery with good discounts. Then came the iPad, third of their iteration in the apple value chain. iPad with retina display promised me the world of color texts, something I missed in my graphite kindle but was not interested to move on to the Kindle HD. I again found myself syncing PDFs of text, book purchases through the Kindle app and DropBox account with the iPad. Yes, its a disruption. But do I still need Paper. Hell Yeah. 
I believe the advances in electronic displays and online publishing has definitely created a dent in the printed text industry. It mirrored the fine nuances of a printed paper, but the feel of holding a book, with the smell of the printed pages is a costly proposition for digital technology. If reading was all about going over a written text, the current state of the art would have been enough. 
I believe we need a hybrid shot at this. A system that allows us the impression of holding a book (and not a library of them) at one time. Similar to what we do now. We need the accessibility, Search and portability of electronic displays and online text, but with the ability of making a real world bookmark, highlighting and marking the written text for notes, and the feel of holding a paper. 
Is it merely a part of the old experience that we crave for or is this a marriage of the real with the virtual world ?
The onset of technologies like Microsoft Kinect with its multi-point Camera, Leap Motion device, feedback touch interaction that can be mashed up to create such an experience. And if this all can be packaged as one small device, well, not sure if it would create the same buzz the same as Google Glass these days.
Paper is not going anywhere, anytime soon. Interactions would evolve and create a hybrid form of human-machine contact. A contact which is human, and yet in its functionality transcends the current limitations of the earlier interaction.

PLEASE NOTE:- The Image used here is a part of the published article "". The blog owner does not own any rights for the image.

Monday, February 04, 2013

One Question to rule them all

Lately I have been interviewing, especially people with technical skills. One question that lingered in my mind for sometime is that of "the" question which would help me get the maximum out of an interview. How about one "big" question, that provides me all the layers of the candidate. It's like an elevator pitch version of interview question.

Out of all the questions I iterated through, I found one. One elegant question. One question to rule them all ;-)

Question - What are three things in your career till now you are proud of ?

Why I believe this is the elegant question :-

1. It proves that the candidate works for purpose

2. It proves the candidate is ambitious, risk taker and values his / her contribution

3. She / he cares for the work

4. Being proud of one's career highlights does not just translate to achievements or successes. It could manifest into learning points, experiences, impacts and influences.

5. The candidate thinks 

6. The candidate is upbeat, optimistic and reflects on the learnings of past.

I think this will be my question of choice in any interview - be it hiring for tech guys, managers, security guards, chefs etc.

What is your question of choice ?

Wednesday, January 30, 2013

Welcome CloudGeek.IN

Launching today, is my new blog "CloudGeek" @

This aims to be a one stop for my thoughts, patterns, architectures and code for the next generation of Cloud Services.

This means, the current blog will be purely personal and will reflect my journey into the unknown. For all my geek readers interested in keeping up with the state-of-the-art in Cloud Computing, grab your seats at

Sunday, January 27, 2013

IaaS giving way to "Data Science as a Service"

As organisations start embracing public infrastructure cloud for their critical data and application needs, Data Science SaaS will become more practical for use. The biggest challenge for Data Science SaaS is to have customers expose / store their data to / on their multi-tenant public platform. Increasingly, public cloud services are becoming a compelling ground for enterprises. This means with applications, data also becomes a good prospect to be moved to these platforms. 
This is due to low data access latency requirements for apps in cloud and increasing confidence in shared public cloud services. 
Once the data is out of the enterprise data center, inter cloud service integration becomes easier. For eg:- Imagine an AWS EC2 infrastructure running your enterprise application, with data lying on S3 / RDS. To extract Business Intelligence and inference out of this data, it is practical to use an existing public SaaS that can work on this data, rather than using in-house analytics infrastructure. Greenfield apps having their genesis on public cloud are already candidate to be used with Data Science services (aka Analytics as a Service). 
The future of Data Science SaaS looks promising. Startups like Datameer, ClearStory are doing pioneering work on this.
More on this at :- 

Tuesday, January 22, 2013

Run in Public, Bring it back when its ready

The road to set up your own cloud infrastructure in your backyard is not a long shot, thanks to the array of techniques / patterns / solutions in the space. However, more organizations are looking at using Public Cloud services, especially IaaS and PaaS to experiment what it takes to scale their apps, and then bring the homework back to their own internal infrastructure a.k.a Z-Cloud of Zynga. This is proving to be a common place for organizations these days. And it makes a lot of sense.
Using public cloud services to test the waters in terms of what it takes to sustain and manage application scale is a good pattern. Once the pattern is understood, and the organization understands the nitty-gritties of managing the scale, an internal infrastructure can be setup, thereby owning the control back. This allows for efficient capacity planning and avoiding the common gotchas of deploying and managing the app at cloud scale.

Monday, January 21, 2013

The rise of Cost aware Architectures

Cost aware architectures have been on rise, thanks to the "as a service" phenomena. A lot have been said around these new evolving architecture patterns. The challenge is now to discover some of patterns that architects can use to build cost-aware architectures. In a gist, a Cost aware architecture is defined as an architecture that evolves based on how it sustains and grows revenues when using price varying infrastructure / application / platform services a.k.a cloud services. This means, as the business gets successful, and starts pooling in more revenue from the user, the cost-dynamics of adding capacity (infrastructure, application and platform) becomes less than linear.
As we see more and more application architectures incorporate "as a service" models, this is going to be the trend for future. I like what Werner Vogels talked about it in an article not long back.


Thursday, January 17, 2013

PaaS is real ? is it... circa 2013

Reposting my earlier post from Jan 2012. How far we have reached with the below predictions ?


I was encouraged to write this post by the influx of articles relating to predictions about Cloud computing in 2012. A common theme surrounding most of them is the rise of PaaS (Platform as a Service). While 2010 and 2011 was spent in understanding the real impact of IaaS and the value it brings, 2012 is destined to be the year when PaaS comes out of the closet with mass acceptance. Following are my ideas on what PaaS would be like in 2012 :-

1. Domain specific PaaS 
A domain specific PaaS would allow organizations of all types to build software infrastructure over existing domain platforms. Lets talk about Telecom domain. There are various IT implementations that need to be built, maintained and supported on a daily basis for a telecom provider. These include provisioning, charging, customer on boarding, analytics, customer service and support, value added services etc. All such application environments need to have a common infrastructure of reusable components ready to be deployed and orchestrated to create solutions.  A Telecom PaaS would provide telcos an unprecedented access to scale, grow and utilize the best practices of the industry without reinventing the wheel. This would create level playing opportunities for telecom providers of all sizes. Imagine how it would affect other verticals - Manufacturing, Egovernance, Logistics, Retail etc. 

2. PaaS enabler buzzword
IaaS platforms would attempt morphing into PaaS enablers. This would allow organizations to create PaaS platforms for the specific business needs at will. Currently the approach the organizations take in this regard : Use existing / Implement an IaaS layer, create application provisioning infrastructure, support deployment of apps on demand etc. However there are challenges that stand before a perfect implementation of PaaS environments using the traditional means. No longer organizations need a bespoke implementation of application provisioning and deployment automation. IaaS players would allow customers to start building these platforms. A lot of traction is already visible in this space. More on this in the next point.

3. Platform-enable anything
The rise of API enabling eco-system to drive the functionalities over a service has been the biggest driver of innovation. Social networks are the living breeding example of such a revolution. PaaS enablers and newer platform services would create opportunities for organizations to platform enable their services at a fast pace. This would allow a faster time to market and easier setup process. Currently this process requires a strategic cap-ex based approach for its implementation. Its not for the weak hearted. A sustainable process to create platforms for service eco-system is the need of the hour. Take example of an Ecommerce enterprise :- With PaaS enablers, the enterprise would be able to create platform for hosting applications built by third-party developers using its services. 

What do you think of this trend ? Share your comments and ideas. 

In the next post, I would like to touch upon Domain specific clouds in more detail. More on this later.