Monday, March 21, 2005
Q & A With Become CTO Yeogirl YunLast week, Become.com sent out a press release that talked about it's patent pending ranking algorithm dubbed AIR. AIR stands for Affinity Index Ranking and based on claims from Become is the next generation search engine ranking algorithm.
Well, we all know how press releases are geared more for investors than techies and how some substance can get lost in the fluff. So I decided to ask Become's Founder and CTO, Yeogirl Yun a few questions to clear things up. The questions are fairly technical in nature but also give a good glimpse into the culture at become from a top down point of view. If you ever wondered what a rockstar of the internet looks like, then look no further. Yeogirl's definitely got the rockstar looks and the coding experience to back it up. Anyway, enjoy the interview.
Q & A With Become CTO Yeogirl Yun
Q: Based on the press release on AIR Become said "AIR identifies exceptional web pages by understanding the level of interconnection between valuable sites from within specific fields of interest. AIR evaluates a web page based on what other "knowledgeable" sites in that specific field say about the page, and also evaluates the page based on what the page says about other "knowledgeable" sites in the specific field." This sounds a lot like the hilltop ranking algorithm whereby known sites are at the top of the hill and they control the ranking landscape by their relationships with other sites lower on the hill. How is AIR different than the hilltop algorithm?
A: AIR is significantly more advanced than that hilltop algorithm. The hilltop algorithm (as described at www.cs.toronto.edu/~georgem/hilltop/) considers only links from a limited number of "expert" sources when identifying target web pages. According to the hilltop paper, "the targets are then ranked according to the number and relevance of non-affiliated experts that point to them. When such a pool of experts is not available, Hilltop provides no results. Thus, Hilltop is tuned for result accuracy and not query coverage."
AIR, on the other hand, evaluates connectivity between all pages in a given topic. Rather than focusing on "top of the hill" sites, AIR understands the overall network of sites within a topical area. Both inlinks and outlinks are evaluated to understand the level of interconnection among the sites. Advanced mathematics and concepts from Applied Physics and Engineering Dynamics are used to calculate specific scores.
Q: AIR technology stands for Affinity Index Ranking and based on the Become.com press release is "ideally suited to specific topical areas such as shopping research". That being said, can AIR only be applied to the shopping vertical or do you see the technology within AIR being applied to several verticals?
A: Although we are focused on the shopping market, AIR is extremely well suited other vertical markets.
Q: I'm curious about this statement "AIR represents a fundamentally new approach to ranking web pages that integrates advanced concepts from Applied Physics and Engineering Dynamics", exactly what concepts from Applied Physics are you referencing in AIR? My brother-in-law's Doctorate is in Physics and is a Professor at Yale, so if I were to ask him how a search engine could benefit from this particular concept in Physics and apply them to their ranking algorithm, what would he say?
A: Great question. At this time, we would like to maintain confidentiality to prevent your brother-in-law from developing a competing engine :) In the future, we plan to publish this information and fully disclose our patent application.
Q: AIR still uses the concept of linking as a pivotal role in evaluating the relationship between sites of importance with regards to ranking them. Do you see any way, in the forseeable future, for a search engine ranking algorithm to completely avoid using linking as one of it's core components?
A: We believe that an understanding of connectivity on the web is an important factor in ranking web pages. We believe, however, that Become's methods of evaluating connectivity are significantly more sophisticated and accurate than the methods used by major search engines today.
Q: When I came to your headquarters for the prelaunch of Become.com I was amazed at the fact you built the entire code base for Become's search engine from scratch in only 9.5 months. I don't honestly know of a single other person on this planet that could accomplish such a feat, it's amazing. How did you do it and what were the critical pieces of your past search related experience [from Wisenut and MySimon] played the most pivotal role?
A: Thank you. We have been extremely pleased with our progress. An important lesson from Wisenut was that search engine development is an extremely difficult challenge. I quickly realized that use of conventional methods and technology would have always kept us several years behind Yahoo and Google. I knew that Become could only succeed by using a completely new approach based on cutting-edge technology and it has been very successful so far. The most important factor in building this new technology has been the hiring of an exceptional team. Every employee at Become has a very strong background and is extremely dedicated to our success. We have also placed high importance on creating a great working environment that encourages people to push the envelope of technical innovation.
Q: You told me about the process you used to screen applicants for positions at Become.com and I found it quite interesting. Would you be willing to share some of that process with my readers? I think it is important because not only have you done incredible things as an individual developer and CTO but you and the Become.com Executive Team have accomplished the almost impossible task of putting together a crack team of employees in a very short period of time. That kind of leadership is hard to find and I think it represents the core of Become.com, not only as a search engine but also as a team.
A: We believe that hiring the best people is a key to our success. We used several methods to identify top talent. Every engineer is invited to take a highly challenging programming test prior to being invited to interview. Only those who excel on the test (which often takes over 40 hours to complete) are invited to interview. Other employees are also tested in their specific job area prior to receiving employment offers. We find that the people drawn to accept these challenges and succeed on the tests are generally a great fit for our team.
We have also worked to create a strong culture of teamwork and contribution. We believe this culture has been a strong reason that great people have joined Become.
Q: Along the same lines of acquiring amazing talent, how on earth did you manage to attract the talent of Jon Glick away from Yahoo? And how did you manage to get another industry veteran, Chris Kermoian, as well?
A: We are extremely fortunate to have a strong management team in place. Chris Kermoian joined Become in May 2004 as our fifth full-time employee. As VP of Product Management, he has been an exceptional leader for all of our Product efforts and defined our systems and processes from inception. Jon Glick is an incredible industry expert who formerly worked for Chris at AltaVista. We are gratified that Jon recognized the potential and vision of our company, and Jon has made an immediate impact as the leader of our upcoming comparison shopping service.
By Jason Dowdell at 11:31 AM | Comments (0)