Razorthink Projects

Following are a few of our projects. You can see the breadth, and variety of technologies brought to bear on these problems - and the level of attention and care required for each.
	Demographics from Realtime Data Although most of our projects are executed in a consulting mode, we sometimes invest in an idea whose time we feel has come. Adonme® happens to be one such (due to be launched in October 2011). Its a model that brings transparency and clean data access to an inherently chaotic industry, outdoor advertising. The site itself involves the extensive usage of AJAX development, including extensive use of Google maps. But the bulk of the technical and analytics work went into the collection and representation of data. Extensive usage of video processing technology and of Bayesian models that integrate multiple forms of demographic data, both static and realtime, gives this business the most complete inventory of assets across India, with the ability to accurately predict dollars or rupees per eyeball for your market segment. Sit in your armchair and run an outdoor campaign in India. Conceptualized and built by Razorthink.
	Building Confidence Sometimes you need to turn a problem on its head to come up with an elegant solution. Virima®'s problem lay in a traditionally over-invested area - trying to map all artifacts and relationships in an IT organization. Standard CMDB-like solutions from large competitors took an enormous amount of time to populate, with data obsolescence setting in before the collection was complete - so much so that a CIO remarked about his CA-built product -- "I trust about 40% of the data in it, and it was populated just 6 months ago.". The approach we came up with was radical: The need for updating data is computed by the system itself. If it feels a piece of data is too old or unreliable, and there are many requests for it, it proactively initiates data updation. Data updates are initiated through multiple, sometimes unusual channels. IM and messaging through near-natural-language interfaces for human data, automated scans where possible. The system maintains a "confidence level" for each piece of data based on multiple characteristics, including age, source, verifiability. The system maintains a full auditable log of every single update to a piece of data, and its "view of the world" can be reconstructed from the log, so its possible to regenerate the view of the world if one historical source is found to be unreliable in retrospect. Visualization originally built using Flex/Actionscript but moved to HTML5/CSS3. Plug it in and the system updates itself in less than 60 days. Great for mergers. Conceptualized by Razorthink.
	Scheduling Chaos Much of supply chain software stops at the level of optimizing master production plans - at this level the problems are still tractable and "modelable" using traditional mechanisms. Getting into plants, however, can be messy - and scheduling is often done using manual processes. Formal models of these, when built, often need to be trashed because of small structural or policy changes in the plant. Our approach to this for Adjility®'s client - a Fortune 50 chemical company - was different: Break up the solution methodology into two parts - a planner that measures and directs the solution algorithm, and a detailed scheduler that faithfully represents the ground realities. Build a perturbation algorithm that intelligently traverses the solution space without repetition. Allow incorporation of additional constraints and policies by customizing the detailed scheduler to each scenario. This approach not only allowed the quick and efficient solution of these problems, but was helpful in providing a mechanism to simulate structural changes and load increases. Usable daily schedules, not just pointers. Conceptualized and built by Razorthink.
	Mobile Skin This was a traditional smartphone Webapp with a twist, our client, a large provider of mobile services wanted to build a skinnable app that could be private labeled. The constraints? (a) The private labelers are picky and would want very custom user experiences (b) A two-week leadtime from initiation of a private labeling request to delivery. Our approach? Architected a configurable model with auto-generation of server and client code. Allowed the quick and easy updating of components by simply cutting and pasting from mockups. Built in validation of configuration. Get a new site up in 2 weeks. Conceptualized and built by Razorthink.