Job available: parse quality at Powerset (a Microsoft company)

Tracy Holloway King tracyhollowayking at gmail.com
Fri Jan 16 16:16:51 UTC 2009


Company: Microsoft
Job Title: Software Development Engineer
Job Category: Software Development
Date Posted: 01/14/2009
		
Job Code: 252003
Location: San Francisco, California, USA


The Powerset team within Live Search is looking for a motivated high
achiever with talents in Natural Language Processing (NLP) and machine
learning to help solve some of the most exciting problems in search.

We are looking for a developer to help improve parse quality for
search applications. The current parser uses primarily symbolic
techniques combined with stochastic methods for improving efficiency
and ranking the parse output. This is a full-time position at Live
Search's Powerset group in San Francisco.

Responsibilities will include: improving stochastic ranking of the
parse output; exploring ways to incorporate additional symbolic or
stochastic techniques into the parser; improving back-off techniques;
and prioritizing improvements to the core grammar rules based on data
analysis. In the longer term, responsibilities may extend to designing
and implementing similar techniques (e.g. stochastic ranking) for the
semantic output and for languages other than English. Work will be
done as part of the natural language engineering team which includes
developers for the morphology, syntax and semantics; it will also
involve close interactions with the natural language tools group.

Required Qualifications:
- MS/PhD in Computer Science or Linguistics (or equivalent industry experience)
- Experience applying machine learning techniques to "deep" representations
- Ability to design and implement features for maximum entropy models
- Programming proficiency (preference for C or C++)
- Background in syntax

Preferred Qualifications:
- Experience in error analysis for NLP systems
- Experience with hybrid symbolic-stochastic systems
- Experience with unification-based formalisms


Applications must be submitted electronically via the Microsoft career
site.  For this position see:

http://members.microsoft.com/careers/search/details.aspx?JobID=B94BE0E1-87F7-481B-BAAC-CCCF058C0523

We look forward to hearing from you!



-- 
Natural Language Engineering
Microsoft Live Search
http://www.linkedin.com/in/tracyhollowayking
+1 415 848-7276



More information about the HPSG-L mailing list