[AlignAI]Al ResearchEngineerEngineeringSeoulYear-roundAlignAlThe easiest useranalytics for yourAl-native product.III Align AIAlignASearch ResultsALOGN-XDMy DashbFindsessionswherethechatbotforgot the user0-=ALTOWNIKin0ALXCNGWDN■:-=Ithinkwehadthisconversation.MyGN FTD###That'sgreatWastdoyoulleeaboutg·WHY IN WING THEIR:...-Aor 1,2023, 11359W3MMar31,2023,244AMWhat is"AlignAl"?AlignAlisaproductanalyticsandevaluationsolutionforLLM-poweredconversationalproducts-thinkGoogleAnalytics forChatGPT.In today'sworld,how humansinteractwithcomputershasevolvedsignificantly.Duringtheweb era,interactionswerebasedonsimpleclicks,whileinthemobileera,swipesandpinchesbecamethenorm.Now,weengagewithcomputersusingnaturallanguage.However,whileusingtheseconversationalproducts,several issuescanarise,suchas:Chatbotsgenerateunfoundedorfalseinformation.Chatbotsfail tounderstanduserinputcorrectly,leadingtostrangeresponses.Chatbotsprovideinconsistentanswerstothesamequestion.AlignAlprovidesasolutionthatempowersproductmanagersof AI-driven systems toanalyzeconversationaldataefficientlyandenhanceproductperformance.Whatwe'vebeenuptoRaised a$3.5Mseedround,ledbyKBInvestmentinJanuary2024LaunchedaSaaSsolutioninOctober2023,reachingProductHunt's#3Productof theDaySecuredaJointBusinessRelationshipwithoneoftheBig4consultingfirmsLaunchedanenterpriseofferingwithcustomon-premisesolutionsEstablishedtheU.S.corporateofficeandtheIndianbranchKeyResponsibilitiesSearchSystemResearchandDevelopmentDesign,train,andoptimizeembeddingmodelsfornaturallanguagesearchBuildadvancedembeddingandindexingsystemsutilizingLLMsDevelopmultilingualsearchsystems,includingEnglishandKoreanSearchModelEvaluationSystemDevelopmentResearchanddesignautomatedlabelingmethodsforsearchqualityevaluationImplementlabelingautomationsystemstoassessthequalityofsearchmodelsRequiredQualificationsPossessabachelor'sdegreeorhigherinAI/MLorequivalentpracticalexperienceDemonstrateproficiencyusingAldevelopmentframeworkssuchasPython,PyTorch,andHuggingfaceExhibit adeepunderstandingof LLMandTransformerarchitecturesBringexperiencetrainingAlmodels(e.g.,LLMfine-tuning,embeddingmodels)Show knowledgeof naturallanguageprocessing(NLP)andinformation retrieval(IR)Maintainaflexiblemindsetforsolvingcomplex problemsPreferredQualificationsEarnedamaster'sordoctorateincomputerscience,artificialintelligence,or a relatedfieldUtilizedVectorDatabases(e.g.,Qdrant,Milvus)inpastprojectsorapplicationsAccumulatedover 3yearsofexperience ina relevantfieldDevelopedlightweightmodelsandmultilingualnaturallanguageprocessingsystemsPublishedacademicpapersorcontributedtoresearchprojects(e.g.,corporateinitiatives,LLM,andAl-relatedopen-sourceprojects)You'llhavefunifYouarepassionateaboutAI technologiesandproductsYouenjoyacultureof mutualgrowththroughcodereviewsYou thrivein rapidlychangingenvironmentsandenjoyquicklyvalidatingdiverseideasYou findsatisfactioninsolvingtechnicalproblemswhilebalancingvarioustrade-offs
