Mayank Lahiri

I am currently a senior software engineer at Amazon Web Services (AWS) in San Francisco.

Prior to Amazon, I was the founder of Iceroad, making a serverless cloud product called Baresoil (now open-source).

Prior to Iceroad, I was a software engineer at Google, working out of San Francisco, California (where I also live). I was part of the Index Quantitative team, which had an open mandate to analyze various Google Search index and indexing mechanisms, and attempt to improve them in various ways. This included things like the Web graph, Pagerank and other signals, indexing, and crawl scheduling.

Prior to Google, I was a research scientist at Facebook in Menlo Park, California. I worked on the Page Insights team, building a system for analytics on Facebook Pages, and then on the Location and Events team, working on data analysis and building various backend systems for an early prototype of Facebook's GPS-based Nearby Friends feature.

Prior to Facebook, I received my Ph.D. in May 2011 from the University of Illinois at Chicago, where I worked with Professor Tanya Y. Berger-Wolf in the Computational Population Biology Laboratory, part of the Computer Science department. You can find my thesis here, and on Amazon, where it is has been categorized under "Agricultural Sciences" for some reason unknown to me, clearly a lonely outlier in some large, popularity-driven data mining process.

Publications

• [pdf] [code+data] M. Lahiri, C. Tantipathananandh, R. Warungu, D.I. Rubenstein, T.Y. Berger-Wolf. Biometric Animal Databases from Field Photographs: Identification of Individual Zebra in the Wild. Proc. of the ACM International Conference on Multimedia Retrieval (ICMR 2011) , Trento, Italy, 2011.
• [pdf] [code+data] M. Lahiri and M. Cebrian. The genetic algorithm as a general diffusion model for social networks. Proc. of the 24th AAAI Conference on Artificial Intelligence (AAAI 2010), Atlanta, Georgia, 2010.
• [pdf] [code] M. Cebrian, M. Lahiri, N. Oliver, A. Pentland. Measuring the Collective Potential of Populations from Dynamic Interaction Data. IEEE Journal of Selected Topics in Signal Processing, Volume 4, Issue 4 (2010), pp. 677-686.
• [pdf] [code] M. Lahiri and T.Y. Berger-Wolf. Periodic subgraph mining in dynamic networks. Knowledge and Information Systems, Volume 24, Issue 3 (2010), p. 467.
• [pdf] [code] M. Lahiri and T.Y. Berger-Wolf. Mining Periodic Behavior in Dynamic Social Networks. Proc. of the 8th IEEE International Conference on Data Mining (ICDM 2008), Pisa, Italy. December 2008.
• [pdf] M. Lahiri, A. Maiya, R. Sulo, Habiba and T.Y. Berger-Wolf. The Impact of Structural Changes on Predictions of Diffusion in Networks. ICDM Workshop on Analysis of Dynamic Networks, Pisa, Italy. December 2008.
• [pdf] M. Lahiri and T.Y. Berger-Wolf. Structure Prediction in Temporal Networks using Frequent Subgraphs. Proc. IEEE CIDM 2007, Honolulu, Hawaii. April 2007.
• [pdf] C.W. Liew and M. Lahiri. Exploration or Convergence? Another Meta-Control Mechanism for GAs. Proc. FLAIRS 2005, Clearwater, Florida. AAAI Press, 2005
• [pdf] A. Dougherty and M. Lahiri. Shape of ammonium chloride dendrite tips at small supersaturation. Journal of Crystal Growth 274 (2005): 233-240 issues 1-2.

Education

• Ph.D. in Computer Science, 2011, University of Illinois at Chicago, Chicago, IL
• B.S. (Hons.) in Computer Science, 2005, Lafayette College, Easton, PA

Peer Review Activities

• Program Committee member: ICWSM 2014, WWW 2014, WSDM 2014, WWW 2013, SocMP 2012 wkshp. (HPDC 2012), Soma 2012 wkshp. (KDD 2012)
• Reviewer/Subreviewer: AAAI 2013, IEEE Intelligent Systems journal in 2012, ACM CSCW 2013, ACM CIKM 2013, ACM CIKM 2012, ACM SIGCOMM 2011, IEEE ICDM 2010.

Other Stuff

• I'm collecting all my writing in one place, and attempting to consolidate my open-source code on Github.
• Some of my side projects: Maxstagram, if Instagram's photo filters aren't edgy enough for you anymore, and Newstandoff, for keeping tabs on the world's newspapers.
• My ongoing list of really twisty, very fun, light-traffic, perfectly banked, breathtaking San Francisco Bay Area sports scenic drives.
• Stripespotter is receiving some attention in the media: coverage in the New Scientist, blog post at io9, Horsetalk (NZ), Wired UK, treehugger, MSNBC, Popular Science, ReadWriteWeb, Fast Company, cnet, more. Also video here, an interview on Science Update, and the crowning achievement of my research career: as a limerick on NPR's Wait Wait Don't Tell Me.
• My submission won 1st place in the 2010-11 UIC Image of Research competition, which judges aesthetic visualizations of research. The entry described an automatic photo identification system for zebras and other striped animals, which is also available for download.
  

  
  Zebraprinting by Mayank Lahiri is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.
  Based on a work at mayank.lahiri.me.
• I just did a second stint at the truly incredible Ol'Pejeta Conservancy in Kenya, collecting data for my zebra photo-ID program. Here's a short video-blog of my time there (in HD!). Read the research paper about it. And also, this is what "field computer science" looks like. YouTube link.

  I have a massive repository (about 31 GB) of high-resolution zebra photos (as well as smaller quantities of other animals). E-mail me if you have a use for them.

• If you like natural things, you should check out my friend Dawn's line of natural body products that she makes herself with locally foraged ingredients..
• How to game Foursquare in 9 lines of Perl.
• If you don't know what satire is, you probably shouldn't read my devastating rebuttal of all complex systems research ever done, which appeared in the proceedings of Carnegie Mellon's highly prestigious and peer-reviewed SigBovik conference.
• I sometimes upload pictures to Flickr.
• A small script for Firefox that enhances Google Scholar by letting you cross-reference citation lists.
• Some general trivial code that's been accumulating over the years.
• My 80-minute Sudoku solver challenge. A newer, faster self-contained version is here.
• My Erdos number is at most 4.

  One sequence is M. Lahiri → T.Y. Berger-Wolf → E. Reingold → A. Fraenkel → P. Erdos