At the Language Technologies Institute at Carnegie Mellon University, my research focus was on human-computer interaction through spoken dialog systems. However, I ended up studying and publishing about crowdsourcing at least as much as I did on language technologies. Check out a book I contributed to to get an idea of what we were working on.
I enjoyed an SDE position at Amazon.com for 4 years, working on data mining and more precisely on information extraction. We powered the faceted search experience of a large portion of Amazon all while helping filling missing catalog values. More recently, I transitioned to a machine learning engineer role with Amazon Music and I have been tackling a few problems including defining good music similarity metrics, genre classification and providing great music recommendation for our customers.
Search engine aggregator to facilitate finding sailboats for sale in your area. Find your dream boat at sailboatlocator.com!
Hosting a party? Set up a decide.fm station and have your hosts decide what song to play next.
Humans adapt to each others at various levels during a conversation. LexE investigates how this phenomenon transposes to human-computer interaction, and how to use it to provide better dialog systems.
VoIP-enabled spoken dialog system based on Olympus architecture. It will be handling a subset of calls to the City of Pittsburgh 311 line.
Crowdsourcing for Speech Processing: Applications to Data Collection, Transcription and Assessment. Wiley, Hoboken, USA, May 2013. [Amazon]
Speaking to the Crowd: looking at the past achievements in using crowdsourcing for speech and predicting future challenges. Interspeech 2011. [PDF]
Spoken Dialog Challenge 2010: Comparison of Live and Control Test Results SIGDIAL 2011. [PDF]
CHI 2011. [PDF] Sources of Variability and Adaptive Tasks.
Toward Better Crowdsourced Transcription: Transcription of a Year of the Let's Go Bus Information System Data. SLT 2010. [PDF]
Spoken Dialog Challenge 2010. SLT 2010. [PDF]
Lexical Entrainment of Real Users in the Let's Go Spoken Dialog System. Interspeech 2010. [PDF]
Clustering dictionary definitions using Amazon Mechanical Turk. NAACL/HLT 2010 MTurk workshop. [PDF]
Multimodal learning of words: A study on the use of speech synthesis to reinforce written text in L2 language learning. SLaTE 2010. [PDF]
Annotation d'expressions temporelles et d'événements en français. TALN2008. [PDF]