Anand Natrajan

Discovery Inc. was a US-based TV network with several popular TV channels in its stable, such as Discovery, Animal Planet, HGTV, TLC, Food Network, EuroSport and others. The company launched a streaming service called discovery+ with an ad-free and ad-supported subscription model.

LucidMedia Inc. was a VC-funded company located in Reston, VA. The company operated in the online advertising space by providing data services and running online campaigns for Fortune 500 customers.

The company achieved unprecedented efficiency in executing ad campaigns primarily because of its flagship technology, ClickSense. Under my leadership, ClickSense evolved from a contextualisation engine to a full-featured ad server and Demand-Side Platform. One of the key functions of this platform was high performance as measured by low latency, high throughput and near-zero downtime.

I was responsible for overall technology direction, including strategy. My responsibilities included product management, partner relationships, technology evangelism and intellectual property management. I was involved daily in the design of the product. My team was also responsible for the deployment and operation of the product, the knowledge base used by the product as well as system infrastructure. I was able to achieve these multiple functions with a small team by emphasising automation of tasks as much as possible, and anticipating use cases as early as possible.

IntelligenX, previously, i411 Inc. was a self-funded, profitable company located in Herndon, VA. The company designed and deployed Internet Yellow Pages capabilities for its customers. The key benefits of the i411 approach were low cost and very high performance.

As a director in a small company, I played a very hands-on role in the design and development of i411 products and strategies. I researched software components that could cooperate or compete with the company's product. I also participated in grant proposal authorship. I worked closely with the product management team to achieve my goals.

I was responsible for the ground-up design of v5.0 of DiscoveryEngine, the company flagship product. This product was the best-of-breed in the faceted, fielded, fulltext search engine space, and was deployed to power a number of customer sites.

The department of Computer Science at the University of Virginia is located in Charlottesville, VA. In the mid-90s, the highly-ranked Computer Science dept. developed a pioneering technology called metasystems (later called grids, now called clouds) under the direction of Prof. Andrew Grimshaw.

The research team, called Legion, was attuned to the needs of several users and collaborators spread across the globe. The development environment was C++, with some Perl/ksh scripting. My focus was on developing quality software in an academic environment and aggressively publishing the results of using this software.

EarthMusic Network was a Hindi film music site located at www.earthmusic.net. The web site provided a unique search engine indispensable to fans of Hindi music. Hindi songs use a Devnagari script (देवनागरी) which is often transliterated into the Roman script for searching and viewing. The transliteration is not canonical — different people transliterate differently. Moreover, information available on Hindi songs is often scant.

I designed and developed a fuzzy search engine, called Mantra, which was incorporated into the general search capability on the site which permitted searching based on singer, film name, year, etc. The web site received 10,000+ hits a day (~ one per 8 seconds), and was one of the most popular web sites for searching for Hindi music.

The topic of my dissertation was "Consistence Maintenance in Concurrent Representations". When independently-constructed models of a certain phenomenon cooperate, their representation of entity states can diverge very quickly, causing the cooperative effort to become useless or even outright wrong. My thesis presented a framework wherein entity states in each model can be made to share information with other models thus ensuring that all of them have a consistent state of the world. I identified the challenges in maintaining this consistency when different types of interactions can change the state within any model.

My Master's research was a prelude that led to my dissertation work. During my Master's I identified the key problem of inconsistent states within multiple models. The publication of my initial work led to a Best Paper Award. Along with an early successful attempt at the Ph.D. qualifiers, this work encouraged me to pursue doctoral studies.

My work was partially funded by the U.S. Department of Defence. I co-authored several of the grants that made funding my own work possible. During my studies, I also took on research and teaching assistantships and mentoring activities. I actively published papers in my research area.