• Share
  • Email
  • Embed
  • Like
  • Save
  • Private Content
Using solr to find the right person for the right job - By Kang Laura
 

Using solr to find the right person for the right job - By Kang Laura

on

  • 972 views

 

Statistics

Views

Total Views
972
Views on SlideShare
972
Embed Views
0

Actions

Likes
0
Downloads
9
Comments
0

0 Embeds 0

No embeds

Accessibility

Categories

Upload Details

Uploaded via as Microsoft PowerPoint

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    Using solr to find the right person for the right job - By Kang Laura   Using solr to find the right person for the right job - By Kang Laura Presentation Transcript

    • Using Solr to find the right person for the right job
      MaY 2011
      Laura Kang
      Theladders.com
    • Agenda
      Search at the Ladders
      Current Projects
      Standardization of search products
      Recommendation service
      Suggested Candidates
      Testing Sessions
      Challenges and Future Work
    • TheLadders.com
      $100K+ job search and career management
      Right person for the right job, right job for the right person
      >4 million members
      Recruit Ladder
      Community of recruitersand job seekers
      Job search advisors and talent specialists
    • Search at the Ladders
      January 2010: search team formed.
      First member: computational linguist, Dr. Leslie Barrett
      February 2011: platform team formed.
      9 members
      Search, CMS, shared services
      Led by Ed Cudahy
    • Solr/Luceneat the Ladders
      Search
      Subscriber search
      Candidate search
      Recruiter search
      Job search
      Hiring
      alerts
      search
      Editorial content
      search
      Suggestions
      Job to
      Candidate
      Candidate to Job
      Recruiter to Candidate
      Recruiters
      like this
      Candidates
      like this
      Jobs
      like this
    • Standardization of Search Products
      Before
      Legacy implementations
      Local Luceneindex
      Different versions of Solr
      Hard to make improvements
      Move to Solr 3.1.0
    • Solr Setup
      Extended dismax
      Preprocessing non-alphanumeric characters
      Analyzers/Tokenizers
      Stop word removal
      Index-side synonym expansion
      Custom stemmers
      Custom length normalization for job description and resumes
      Replication
      Multi-core
      Template: maven archetype
      Autosuggest
    • Challenges
      Power users
      Large Boolean queries
      Target result set of 50
      Synonyms with different lengths
      HR manager vs. Human Resource manager
      Phrase slop
      Company variants
      Communicating to users
    • Typical Query
      Title: "sales consultant" "sales engineer" "solutions architect" "solution architect" engineer architect "pre-sales consultant" "presales consultant" "solutions engineer" -president -vp -cmo -ceo -cfo -chief -director -software -"s/w" –database
      Location: Redwood City, CA, 100 miles
      Company: apple "ingram micro" "tech data" intermec "insight enterprises" "super micro computer" "digi international inc" radisys "silicon graphics international corp" crayangilysysvoltaire "concurrent computer" gtsi "socket mobile" hphewletthitachiibm "i. b. m." "i.b.m." "international business machines" "business machines" intevacxyratex quantum "western digital" wdw.d. ramtron "micron technology" e.m.c. "e. m. c." sandisk brocade seagate 3par stec "dot hill" oczo.c.z. "o. c. z." "hutchinson technology" "hutchinson tech" lasercardedcidataram "overland storage" emcnetapp
      Keyword: (presales "pre sales" pre-sales sales) AND (engineer engineering) AND (storage hardware "sun microsystems" "sun servers")
    • Recommendation Service
      RESTful web service
      Solr backend
      Tracking/feedback
      A/B framework
      Client Application
      Recommendation Service
      Job to
      Candidate
      Candidate to Job
      Recruiter to Candidate
      Recruiters
      like this
      Candidates
      like this
      Jobs
      like this
    • Other Projects
      Resume parsing
      Parse flagging
      Company list
      Company variants
      Similar companies
      Synonym list generation
      Search help for users
    • Suggested Candidates
      Coming soon on Recruit Ladder
      http://recruit.theladders.com
      Suggest candidates that might be a good match when a recruiter posts a job
      Goal:
      Help recruiters quickly identify candidates that are right for the job
      Increase interaction between recruiters and job seekers
    • Suggested Candidates
      Hybrid
      Categories
      Position level
      Job specialties/disciplines
      Industry
      Job title => position level (VP, director) + job function (marketing)
      Relevance matrices for partial matches
      Text
      Job function keywords
      Candidate’s job experience vs. job description using MoreLikeThis
      Boosting
      Filters: salary range, years of experience, location
    • Challenges
      Certifications and skills
      Taxonomy
      “Good” resume
      Resume scoring
      Job description
      Short job descriptions
      “Series 7 & 63 and 10+ years of experience in related field required”
      Company description/Equal Employment Opportunity clauses
      Different priorities for each job type
    • Evaluation
      Other products:
      Mean-average precision: search
      A/B testing: suggested jobs
      Testing sessions with talent specialists and external recruiters
      Randomly selected set of jobs
      Given 5 candidates and their resumes
      0 = not a good match
      1 = keep in pipeline
      2 = contact
    • Future Work
      Job description parser
      Skills and certifications
      UI improvements
      Personalized weights and filters
      Company description
      Geospatial features
      Recommendation Service: user ratings
      Search/Recommendation quality testing framework
      Customer satisfaction metrics
      Pre-release metrics
    • We’re hiring
      Email lkang@theladders.com
    • Questions?