Personal Information
Entreprise/Lieu de travail
San Francisco Bay Area United States
Profession
Engineering at LinkedIn
Site Web
www.linkedin.com
À propos
Objective: Engineer systems & algorithms to help users get to the content they need.
Summary:
Hands-on experience with distributed systems for both online and offline data processing.
Designed and implemented low-latency high-throughput online retrieval systems from scratch, doing micro and millisecond latencies for few hundred QPS per node (without caching).
Designed and implemented simple & extensible data-infrastructure for offline data processing pipelines on hadoop. These range from simple search-index building pipelines, to non-trivial pipelines to do machine learning algorithms. Using tools like plain java map/reduce, pig, hive, spark, scalding and so forth (ordered by familiari...
Mots-clés
concurrency
jvm
stm
multi-core
lock-free
transactional memory
overview
presto
mapreduce
hadoop
hdfs
spark
big data
Tout plus
Présentations
(2)J’aime
(3)Invokedynamic in 45 Minutes
Charles Nutter
•
il y a 11 ans
Distributed Consensus A.K.A. "What do we eat for lunch?"
Konrad Malawski
•
il y a 9 ans
DocValues aka. Column Stride Fields in Lucene 4.0 - By Willnauer Simon
lucenerevolution
•
il y a 12 ans
Personal Information
Entreprise/Lieu de travail
San Francisco Bay Area United States
Profession
Engineering at LinkedIn
Site Web
www.linkedin.com
À propos
Objective: Engineer systems & algorithms to help users get to the content they need.
Summary:
Hands-on experience with distributed systems for both online and offline data processing.
Designed and implemented low-latency high-throughput online retrieval systems from scratch, doing micro and millisecond latencies for few hundred QPS per node (without caching).
Designed and implemented simple & extensible data-infrastructure for offline data processing pipelines on hadoop. These range from simple search-index building pipelines, to non-trivial pipelines to do machine learning algorithms. Using tools like plain java map/reduce, pig, hive, spark, scalding and so forth (ordered by familiari...
Mots-clés
concurrency
jvm
stm
multi-core
lock-free
transactional memory
overview
presto
mapreduce
hadoop
hdfs
spark
big data
Tout plus