AIResearchEngineer위치서울팀Motif TechnologiesAbouttheRoleAsanResearchEngineer, you willbe at theforefrontofinnovation,leadingthedevelopment ofnext-generationgenerativeAImodelsandtechnologies acrossdiverse domains,includingLLMs,Text-to-Image,andText-to-Video.Your work willdrivetheevolution of AI,blendinggroundbreaking researchwith real-worldimpact,andtransformingcutting-edgeideasintotransformative solutionsthat shapethefuture.This is more than a job-it's aninvitationtopushtheboundaries of what's pssible.You'll immerseyourself in solvingsomeofAl'smostchallenging problems,all whilecreating productsthat redefineindustriesandinspirenewpossibilities.We fosteranenvironmentthatthrivesoncuriosity,creativity,andfearlessexperimentation,empowering youto turn boldideasintoreality.If you're passionateaboutadvancingthestateof theart in Al,driven tocreatemeaningfulimpact,andeagertobuildthefutureof Al-powered solutions,this roleisyouropportunity toshine.Join us,and bepartofateamwhereinnovation knows nolimits.ScopeofResponsibilitiesPre/Mid/Post-trainingInferenceScaling LawsCoT/ReasoningDataprocessingEvaluationAlignmentArchitectureDistributedtrainingLLM/Image/Video/AudioMLOpsYouwill(Architecture,LLM/Image/Video/Audio)Conductappliedresearch foradvancing neuralnetwork architecturesof flagshiplanguage/image/video/etcmodels.(Evaluation,LLM/Image/Video/Audio,Dataprocessing)Conductappliedresearchonevaluationcriteriaforflagshiplanguage/image/video/etcmodels.(Pre/Mid/Post-training,Distributedtraining,Dataprocessing,LLM/Image/Video/Audio)Conductapplied research onlarge-scaletraining/efficientsampling/optimization.(Pre/Mid/Post-training,Distributedtraining,ScalingLaws)Conductappliedresearchforimprovingthetrainingthroughput.(Data processing,Evaluation,Pre/Mid/Post-training)Conduct appliedresearch forcollectingandevaluating new metods forhighquality data.(Evaluation,CoT/Reasoning,Pre/Mid/Post-training,LLM/Image/Video/Audio,Alignment,Dataprocessing)Conductappliedresearch toimprovetheabilityof foundationalmodelstoaccurately reasonaboutquestions ofhumanvalues,morals,ethics,andculturalnorms,andapplytheseimproved models topracticalsafetychallenges.(Inference,LLM/Image/Video/Audio)Conductappliedresearch on generativemodelarchitecture tomakethem moresuitableforefficientinference.(MLOps,Dataprocessing)Design,build,andoperateMLOpssystemsthat can help youefficientlytrain,infer,andevaluatemodels.[지원하기]
