Blog aggregator similar to techmeme

I am looking to develop a platform similar to [login to view URL] but for a different industry. I have a collection of approximately 1200 blogs that can be used to seed. The site would behave much like techmeme in that it would 1) scrape websites/blogs for data 2) collect and index the data 3) Algorithmically or by way of machine learning cluster articles/posts that relate to each other 4) present the data in real time in a structured and easy to navigate way 5) provide a backend that would allow an administrator/user some "editorializing" such as tagging one article/post in a cluster as the top story. Backend also needs to be able to manage all aspects of website - sponsorship, users, database updates, add new urls to the scraper, etc. 6) provide a means to organize and promote sponsorships throughout the site. Based on my research, this project could be accomplished using a combination of Apache Nutch, solr, hadoop, and mahout. This will likely be deployed on a platform like Amazon AWS. Type of Website: News Media / Informational Content Other Skills: hadoop, mahout, nutch, solr, lucene, java