Turing会社紹介資料

by Turing

Embed

Start on current slide

Slide 1

Slide 1 text

Slide 2

Slide 2 text

2 ೔ຊͷࣗಈं࢈ۀ͸ GDP ͷ໿1ׂʹ50ஹԁΛ઎ΊΔࠃՈͷجװ࢈ۀͩɻ 80೥௒ͷྺ࢙ΛތΔτϤλ͕௕Β͘ੈքटҐΛक͖͕ͬͯͨɺ ถதͷελʔτΞοϓ͕2000೥୅Ҏ߱ɺӍޙͷ᝔ͷΑ͏ʹ୆಄͠ɺͦͷԦ࠲Λڴ͔͍ͯ͠Δɻ ͍·ۀք͸“ήʔϜνΣϯδ”ͷͨͩதʹ͋Γɺ͔ͩΒͦ͜௅ΉՁ஋͕͋Δɻ ৼΓฦΕ͹1989೥ɺੈք࣌Ձ૯ֹϥϯΩϯάTOP50ͷ͏ͪ32ࣾΛ೔ຊاۀ͕઎Ί͍ͯͨɻ ͔͠͠ݱࡏɺ࢒Δͷ͸Θ͔ͣ1ࣾɻ্Ґ͸΄΅ถࠃاۀʹಠ઎͞Ε͍ͯΔɻ ͦͷࠩΛੜΜͩ΋ͷ͸ɺੈքΛม͑Δͱ“צҧ͍”ͯ͠ಈ͖ग़͢ਓͷଟͩ͞Ζ͏ɻ “We Overtake Tesla” ΛॳΊͯޱʹग़ͨ࣌͠͸ා͔ͬͨɻ໌Β͔ʹצҧ͍໺࿠͔ͩΒͩɻ Ͱ΋ޱʹग़ͯ͠े෼ͳਓ͸͓ͦΒ͍͘ͳ͍ɻ͔ͩΒͦ͜ޱʹग़ͦ͏ͱࢥͬͨɻ ࢲͨͪҎ֎ͷͨ͘͞Μͷਓ΋צҧ͍Λͯ͘͠ΕΔ͜ͱΛفͬͯେ੠Ͱצҧ͍Λએݴ͢Δɻ ”We Overtake Tesla”ͱ͍͏੤͍ ϝοηʔδ TURING COMPANY DECK | MESSAGE

Slide 3

Slide 3 text

3 MISSION ׬શࣗಈӡసΛ࣮ݱ͢Δ TURING COMPANY DECK | MISSION

Slide 4

Slide 4 text

4 Technology ։ൃ֓ཁ TURING COMPANY DECK | PROJECT

Slide 5

Slide 5 text

5 TURING COMPANY DECK | Technology ैདྷͷγεςϜ: ಠཱͳϞδϡʔϧͰߏ੒ End-to-Endࣗಈӡసͷ֓ཁ End-to-EndϞσϧ VLAϞσϧ ը૾ LiDAR఺܈ HDϚοϓ Perception ●෺ମೝࣝ ●ඪࣝೝࣝ ●Ϩʔϯೝࣝ Prediction ●Ҡಈ༧ଌ ●কདྷϚοϓ༧ଌ ●ަ௨ΤʔδΣϯτ Planning ●୳ࡧ໰୊ ●ܦ࿏ܭը Control ●੍ޚΞϧΰϦ ζϜ զʑͷΞϓϩʔν: End-to-endࣗಈӡస ϚϧνΧϝϥը૾ ंͷܦ࿏

Slide 6

Slide 6 text

6 ͜Ε·Ͱͷ։ൃਐḿ TURING COMPANY DECK | PROJECT ૑ۀ1೥൒Ͱɺࠃ಺ൢചं྆ͷADAS※ ͱಉ౳Ҏ্ͷੑೳͷࢧԉػೳΛ։ൃɻ  ࣍͸Door to DoorͷࣗಈӡసγεςϜΛ։ൃ͢Δɻ ެಓͰͷࣗಈӡస ౦ژ౎಺Λ30෼Ҏ্  ਓؒͷհೖͳ͠Ͱ૸ߦ͢Δ ެಓͰͷӡసࢧԉ  ࠃ಺ൢചं྆ͱಉ౳Ҏ্ͷੑೳΛ  ࣋ͬͨӡసࢧԉػೳ ࣗಈӡసϨϕϧ2ͷ ADASͱͯ͠  ࢢൢंʹ౥ࡌ͠ൢച ※ADASʹAdvanced driver-assistance systemsࣗಈंͷυϥΠόʔͷӡసૢ ࡞Λࢧԉ͢ΔγεςϜɻ 2022.01 2023.01 2025.11 ࢲ༗஍Ͱͷपճӡస  ςετίʔε಺ͷ  ܾΊΒΕͨϧʔτΛࣗಈӡస

Slide 7

Slide 7 text

TURING COMPANY DECK | Technology Tokyo30

Slide 8

Slide 8 text

TURING COMPANY DECK | Technology ࣗಈӡసAIͷਐԽͱ2030೥ͷల๬ 2020೥ʙ ʙ2020೥ ं྆ VLAϞσϧ ӡసੑೳʹՃ͑ͯɺݴޠཧղೳྗΛ΋ ͪɺ͋ΒΏΔަ௨ঢ়گʹରԠ͢Δɻ Χϝϥ 2030೥ʙ ϑΟδΧϧج൫Ϟσϧ L4(ODD֦େ) / L5 ं྆ ࣗಈӡసAIϞσϧ ୯ҰͷAIͰೝࣝɾܭըɾ࣮ߦΛ࣮ݱɻ  ެಓʹ͓͍ͯਓΑΓߴ͍ೳྗΛ΋ͭɻ Χϝϥ 2024೥ʙ End-to-End AI L2+(ࢢ֗஍) / L4 ※E2E ं྆ Χϝϥ Lidar HD Map ܦ࿏൑அAIϞσϧ ೝࣝAIϞσϧ ϧʔϧϕʔεܕ  ࣗಈӡస ܦ࿏൑அAI ࣄྫ L4 ଟ͍ / ؆୯ গͳ͍ / ೉͍͠ ೉͠͞ ద༻  ൣғ L2ߴ଎ಓ࿏

Slide 9

Slide 9 text

9 ֤։ൃνʔϜͱ࿈ಈ͢Δɺ։ൃίϯϙʔωϯτ ։ൃͷશମ֓ཁ URING COMPANY DECK | Team Driving System 1/2 Driving AI 1 ج൫AI Πϯϑϥ Hardware OS, Middleware Application AI Χϝϥ ΞϓϦ ܭࢉث ։ൃνʔϜ Driving AI 2 Driving AI 3 MLOps 1 MLOps 2

Slide 10

Slide 10 text

DriveHeron 10 ։ൃ಺༰ TURING COMPANY DECK | Team ج൫AI νʔϜ ۩ମతʹ͸ ● ࢹ֮-ݴޠ-ߦಈ(VLA)Ϟσϧͷߏங ● VLAϞσϧͷ࠷దԽ ● ϞσϧධՁϕϯνϚʔΫࢦඪͷߏங 1.׬શࣗಈӡసʹ޲͚ͨେن໛ج൫Ϟ σϧͷ։ൃ 2.VLAϞσϧͷ࣮ं྆σϓϩΠ Vision-Language Model ϢʔβʔίϚϯυ ं྆ΩωϚςΟΫε ࣌ܥྻͷը૾ϑϨʔϜ αϒλεΫϔο υ Action  ϔου ࣍ͷΞΫγϣϯ αϒλεΫग़ྗ DriveHeron ϞσϧධՁσʔληοτߏங

Slide 11

Slide 11 text

11 TURING COMPANY DECK | Team Driving AI 1 νʔϜ ۩ମతʹ͸ ● End-to-EndࣗಈӡసͷͨΊͷMLϞσϧ։ൃ ● Ϟσϧɾֶशσʔλɾ੍ޚͷվྑ ● σʔλͷ඼࣭վળʢηϯαΩϟϦϒϨʔγϣϯɺࣗݾҐ ஔਪఆɺ࣌ࠁಉظɺetcʣ ● ं੍྆ޚɺ࢟੎ਪఆΞϧΰϦζϜͷ։ൃ ● MLϞσϧධՁͷͨΊͷ࣮ं૸ߦࢼݧ 1.End-to-EndࣗಈӡసϞσϧͷ։ൃ 2.৽نϞσϧΞʔΩςΫνϟɾ৽نं ྆ͷ୳ࡧ ΫϩʔζυίʔεͰͷ੍ޚࢼݧ ࣍ظࣗಈӡసं྆ͷݕ౼ E2EϞσϧͷ૸ߦࢼݧʢ͇̐ഒ଎ʣ

Slide 12

Slide 12 text

12 TURING COMPANY DECK | Team Driving AI 2 νʔϜ ۩ମతʹ͸ ● End-to-EndࣗಈӡసͷͨΊͷMLϞσϧ։ൃ ● σʔλऩूɺσʔλΩϡϨʔγϣϯʹΑΔ࠷దͳσʔλ ηοτ࡞੒ɺσʔλͷεέʔϧΞοϓ ● σʔλͷ඼࣭վળʢηϯαΩϟϦϒϨʔγϣϯɺࣗݾҐ ஔਪఆɺ࣌ࠁಉظɺetcʣ ● ֶशαΠΫϧߴ଎Խʹ޲͚ͨγφϦΦςετઃܭɺσʔ ληοτ࡞੒αΠΫϧͷվળ ● MLϞσϧධՁͷͨΊͷ࣮ं૸ߦࢼݧ ● ࣮ंͰਪ࿦ͤ͞ΔͨΊͷਪ࿦ߴ଎Խ Neural Network 3D෺ମೝࣝ & Ҡಈ༧ଌ Ϛοϓೝࣝ BEV 3࣍ݩಛ௃ multi-camera images ܦ࿏ੜ੒ future trajectory ։ൃ಺༰ 1.End-to-EndࣗಈӡసϞσϧͷ։ൃ 2.Ϟσϧͷෳ਺ंछల։ͷख๏։ൃ 3.ϞσϧͷධՁ͓ΑͼධՁγεςϜͷ վળ

Slide 13

Slide 13 text

13 TURING COMPANY DECK | Team Driving AI 3 νʔϜ ۩ମతʹ͸ ● End-to-EndࣗಈӡసͷͨΊͷMLϞσϧ։ൃ ● େن໛ͳಈըσʔλ׆༻ͨ͠൚Խੑೳͷߴ͍ ੈքϞσϧͷ։ൃ ● ੈքϞσϧͱߦಈϞσϧͷ౷߹͠ߴ඼࣭ͳ૸ ߦσʔλͰFine-tuningʢWAMʣ ● ࣮ंਪ࿦Λݟਾ͑ͨৠཹ ● ਪ࿦ͷߴ଎Խɾ࠷దԽ ։ൃ಺༰ 1.End-to-EndࣗಈӡసϞσϧͷ։ൃ 2.World-Action-Modelͷ։ൃ 3.࣮ंσϓϩΠ ಈըੜ੒ϕʔεͷੈքϞσϧ։ൃ World-Action-Model (WAM) ͷ։ൃ ૸ߦσʔλΛ࢖ͬͨ Fine-tuning ं྆Ͱͷਪ࿦ v0 a0 v1 a1 v2 a2 v3 a3 Context ੜ੒ v 0 a 0 v n a n Context (kv cache) a n+1 ΞΫγϣϯੜ ੒

Slide 14

Slide 14 text

14 TURING COMPANY DECK | Team MLOps 1 νʔϜ ۩ମతʹ͸ ● σʔλج൫ͷ։ൃɾӡ༻ʢS3ɺDatabricks౳ʣ ● Ϟσϧֶशج൫ͷ։ൃɾӡ༻ʢAWS Parallel Cluster ౳ʣ ● σʔλύΠϓϥΠϯͷ։ൃɾӡ༻ʢDatabricks౳ʣ ● ධՁج൫ͷߏங ● ࣾ಺޲͚αʔϏεͷ։ൃɾӡ༻  ྫɿσʔληοτ࡞੒αʔϏεɺσʔλΩϡ ϨʔγϣϯɾՄࢹԽαʔϏε౳ʢAWS API gatewayɺ LambdaɺBatchɺStep functionsɺVercelɺSupabase౳ʣ ● ֤ج൫ɺαʔϏεͷϝτϦΫεऩूͱ؂ࢹʢDatabricks dashboardɺSentry౳ʣ ։ൃ಺༰ 1.ࣗಈӡసϞσϧͷ։ൃޮ཰Λ޲্͞ ͤΔϓϥοτϑΥʔϜɾαʔϏεͷ ։ൃɾӡ༻ MLOpsνʔϜͷΧόʔ͢ΔྖҬ σʔλͷΩϡϨʔγϣϯ

Slide 15

Slide 15 text

15 TURING COMPANY DECK | Team MLOps 2νʔϜ ۩ମతʹ͸ ● γϛϡϨʔλΛ༻͍ͨࣗಈӡసϞσϧධՁύ ΠϓϥΠϯͷ։ൃ(AWS Step Functions, AWS EKS etc…) ● ࣗಈӡసϞσϧ࣮ं૸ߦ࣮ݧͷධՁϝτϦΫ εͷ։ൃ ● E2Eࣗಈӡస޲͚γϛϡϨʔλͷ։ൃ ● ධՁμογϡϘʔυͷ։ൃ ։ൃ಺༰ 1.ࣗಈӡసϞσϧධՁج൫ͷ։ൃ/ӡ༻ 2.3DGSϕʔεͷΫϩʔζυϧʔϓγ ϛϡϨʔλ։ൃ 3DGSΫϩʔζυγϛϡϨʔλ ධՁμογϡϘʔυ

Slide 16

Slide 16 text

16 TURING COMPANY DECK | Team ΠϯϑϥνʔϜ ۩ମతʹ͸ ● GPUΫϥ΢υΠϯϑϥͷ৽نௐୡ/ߏங/ӡ༻ ● ࣗಈӡసAI։ൃʹ࠷దԽͨ͠ܭࢉج൫ઃܭ ● Ϟσϧֶश΍σʔλॲཧʹ͓͍ͯɺ௿ਫ਼౓ԋ ࢉͳͲΛར༻ͯ͠࠷దԽͱߴ଎Խ࣮૷ ● NVIDIA H100 96جͷࣗࣾGPUΫϥελ ʮGaggle Clusterʯ ߏங/ӡ༻ ● ΠϯϑϥՄࢹੑ޲্ ։ൃ಺༰ 1. େن໛ܭࢉج൫ͷӡ༻ɾ੔උ 2. ࣍ظઐ༗Ϋϥελͷݕ౼/ௐୡ 3. ϓϩηεશମͷϘτϧωοΫղফ

Slide 17

Slide 17 text

17 TURING COMPANY DECK | Team Driving System 1/2νʔϜ ۩ମతʹ͸ ● ૊ΈࠐΈLinuxͰ࣮ߦ͢ΔࣗಈӡసγεςϜ։ൃ ● ηϯα੍ޚ΍௨৴༻ϛυϧ΢ΣΞ։ൃ ● ౥ࡌ͢Δϋʔυ΢ΣΞͷબఆɾධՁ ● AIϞσϧͷσϓϩΠ(CUDA, TensorRT) ● σʔλऩूγεςϜͷ։ൃ/ӡ༻ ● ࣗಈӡసͷσϞ༻΍σόοά༻ͷUI։ൃ ● ෳ਺SoC΍BSP޲͚ʹγεςϜҠ২&ݕূ  ͳͲ ։ൃ಺༰ʢେ͖͘3఺ʣ 1.ࣗಈӡసγεςϜͷ։ൃ/ӡ༻  (ιϑτ΢ΣΞ։ൃ, ηϯαʔ੍ޚ) 2.࣮ݧं྆ͷ։ൃ 3.σʔλऩू΍૸ߦ࣮ݧӡ༻ͷ࠷దԽ Χϝϥೖྗ AI ं੍྆ޚ

Slide 18

Slide 18 text

18 TURING COMPANY DECK | Team ڭࢣσʔλͷ४උɺϞσϧઃܭɺֶशɺݕূ·ͰϫϯετοϓͰ MLOpsαΠΫϧΛ௨ͯ͡E2EࣗಈӡసΛ࣮ݱ σʔλऩू ࣮૸ߦͰͷϞσϧݕূ MLOps 18 Ϟσϧ։ൃ ֶशσʔλͷબ୒ɺΫϦʔϯΞοϓɺ  Ξϊςʔγϣϯ

Slide 19

Slide 19 text

19 BUSINESS ࣄۀ TURING COMPANY DECK | BUSINESS

Slide 20

Slide 20 text

20 TURING COMPANY DECK | BUSINESS iPhoneొ৔Ҏ߱ɺ௨৴ɾίϯςϯπ࢈ۀ͕ܹม GAFAMͷ୆಄ɺSNSීٴɺΞϓϦܦࡁݍͷരൃత੒௕ ଟ͘ͷ࢈ۀΛר͖ࠐΈ”ੜ׆༷ࣜ”ͦͷ΋ͷΛม͑ͨ OTAΞοϓσʔτͰंΛৗʹਐԽͤ͞ΔʮεϚϗͷΑ͏ͳंʯ ौ଺θϩɺࣄނθϩɺҠಈࣗ༝ɺ෺ྲྀޮ཰Խ ंपลͷڊେαʔϏεϓϥοτϑΥʔϜͱ৽αʔϏεͷ஀ੜ ׬શࣗಈӡస͕৽ͨͳ  ܦࡁݍɾڊେαʔϏεϓϥοτϑΥʔϜΛੜΈग़͢ ࣗಈӡస͸εϚϗֵ໋Ҏདྷͷେมֵ

Slide 21

Slide 21 text

21 AI͕ະདྷͷंͷίΞٕज़ ׬શࣗಈӡసAI͸εϚϗҎ্ʹਓྨͷੜ׆༷ࣜΛେ ͖͘ม͑Δٕज़Ͱ͋Γɺ͍ۙকདྷʹ࣮ݱ͢Δͱ֬৴ ͍ͯ͠·͢ɻ ׬શࣗಈӡసAI͕࣮ݱͨ͠ະདྷʹɺνϡʔϦϯά͕ ඪ४తͳબ୒ࢶʹͳ͍ͬͯΔΑ͏զʑ͸ࠓ͜ͷॠؒ Λશ଎ྗͰ૸͍ͬͯ·͢ɻ ͋ΒΏΔϝʔΧʔ͕εϚϗతͳंΛ๬Ήͱ͖  νϡʔϦϯάͷAIΛඪ४తͳબ୒ࢶʹ νϡʔϦϯάͷ׬શࣗಈӡసAIΛσϑΝΫτελϯμʔυʹ TURING COMPANY DECK | BUSINESS

Slide 22

Slide 22 text

22 தظతͳϏδωεϚΠϧετʔϯ TURING COMPANY DECK | BUSINESS ܗଶ: ಛఆंछͰಛఆ ٕज़ͷςετ ઌߦ։ൃ ϑΣʔζ 1 ܗଶ: ྔ࢈ं౥ࡌͷͨ Ίͷςετ ઌߦ։ൃ ϑΣʔζ 2 ܗଶ: ྔ࢈ࢼ࡞Λܦͯɹ ྔ࢈ंʹ౥ࡌ ྔ࢈ंʹ౥ࡌ ݱࡏ஍ 2025 2026 2027 2029 ܗଶ: ଟ༷ͳܗଶ PoC / Demo

Slide 23

Slide 23 text

2025೥11݄࣮ূ࣮ݧ 2025೥·Ͱʹ౦ژͷෳࡶͳಓ࿏Λ  30෼Ҏ্հೖͳ͠ͰӡసͰ͖ΔΑ͏ͳɺ  E2EࣗಈӡసϞσϧΛ։ൃ͢Δ TURING COMPANY DECK | PROJECT Tokyo30ୡ੒

Slide 24

Slide 24 text

׬੒ंϝʔΧʔ΍෦඼αϓϥΠϠʔͱڞಉݚڀΛਪਐ TURING COMPANY DECK | BUSINESS কདྷͷE2EϞσϧྔ࢈Λݟਾ͑ɺ  ·ͣ͸ࣗಈंϝʔΧʔ΍෦඼αϓϥΠϠʔͱͷڞಉݚڀ΍ٕज़ఏڙܖ໿Λ໛ࡧ 24 MLOps σʔλऩू Ξϊςʔγϣϯ  ɾֶश ࣮૸ߦͰͷ  Ϟσϧݕূ Ϟσϧ։ൃ E2E MLOps Platform ڭࢣσʔλͷ४උɺϞσϧઃܭɺֶशɺݕূΛ  ࣮ࢪ͢ΔϓϥοτϑΥʔϜج൫Λఏڙ͢Δ ٕज़ఏڙ ڞಉݚڀ ڞಉ։ൃͨ͠ं྆ΛධՁɾվળ ύʔτφʔʢOEM,αϓϥΠϠʔʣͷं྆ʹ  E2EϞσϧΛద߹։ൃ͠ɺੑೳΛධՁ͢Δ ύʔτφʔं྆ E2EϞσϧ

Slide 25

Slide 25 text

νϡʔϦϯά͸ઈ޷ͷҐஔʹɺઈ޷ͷλΠϛϯάͰࣄۀΛల։͍ͯ͠Δ TURING COMPANY DECK | BUSINESS ੓෎ิॿ

Slide 26

Slide 26 text

TURING COMPANY DECK | BUSINESS ͳͥTuring͕ࣗಈӡసͰউͯΔͷ͔ʁ 02 AIΤϯδχΞΛऒ͖͚ͭΔʹ͸ັྗతͳ໨ඪͱ ͦΕʹର͢Δਖ਼͍͠Ξϓϩʔνɺͦͯ͠Ϧʔ μʔγοϓ͕ඞཁɻ Turing͸AI։ൃͰ࣮੷Λ࣋ͭࢁຊCEOͷϦʔ μʔγοϓͷԼɺ׬શࣗಈӡసͱ͍͏ਓྨͷά ϥϯυνϟϨϯδʹ௅Ήɻ ׬શࣗಈӡస͸AIιϑτ΢ΣΞΛ։ൃ͢Δ͚ͩ Ͱ͸࣮ݱͰ͖ͳ͍ɻ ϋʔυ΢ΣΞͱͷ༥߹͕ඞཁෆՄܽͰ͋Γɺߴ ͍ٕज़ྗͱڧݻͳαϓϥΠνΣʔϯΛ࣋ͭ೔ຊ ͔ͩΒͦ͜׬શࣗಈӡసͷ։ൃʹνϟϨϯδͰ ͖Δɻ ࣗಈӡస͸Not a Winner Takes ALL. ೔ຊͷڧ͍ࣗಈं࢈ۀ 01 AI։ൃʹ͓͚ΔϦʔμʔγοϓ 26

Slide 27

Slide 27 text

27 BUSINESS ձࣾ֓ཁɾ૊৫ TURING COMPANY DECK | BUSINESS

Slide 28

Slide 28 text

28 ձࣾ֓ཁ ※2024೥8݄࣌఺ ໊শ Turingגࣜձࣾ ૑ۀ 2021೥8݄20೔ ࣄۀ಺༰ ׬શࣗಈӡసAIͷ։ൃ ຊࣾॴࡏ஍ ౦ژ౎඼઒۠େ࡚1ஸ໨11−2   ήʔτγςΟେ࡚ Πʔετ౩4֊ ୅දऔక໾ ࢁຊ Ұ੒ ࢿຊۚ 3000ສԁʢྦྷܭ240ԯԁௐୡʣ ࣾһ਺ 92໊ ʢ໾һɾਖ਼ࣾһ75໊ɺΠϯλʔϯɾ༗ظޏ༻17໊ʣ TURING COMPANY DECK | COMPANYPROFILE COMPANY

Slide 29

Slide 29 text

૊৫ߏ੒ TURING COMPANY DECK | ORGANIZATION 29

Slide 30

Slide 30 text

૊৫ߏ੒ TURING COMPANY DECK | ORGANIZATION 30 75ਓ ໾һɾਖ਼ࣾһ਺ ૊৫ߏ੒ Seed Pre-Series A Series A 1st close

Slide 31

Slide 31 text

31 ਓһߏ੒ TURING COMPANY DECK | ORGANIZATION ೥ྸ૚ ύύɾϚϚൺ཰ த్ɾ৽ଔ

Slide 32

Slide 32 text

32 Ϛωδϝϯτϝϯόʔ TURING COMPANY DECK | ORGANIZATION ༷ʑͳ෼໺Ͱ৽نࣄۀϓϩδΣ ΫτΛଟ਺ओಋ ΠϯυͰʮYugo SushiʯΛ૑ۀ ͠ɺ੒ޭཪʹӡӦ ελʔτΞοϓCFOͱͯ͠໿220 ԯԁΛௐୡ MS Menlo Parkʹͯྦྷܭ4ஹԁͷ औҾܦݧΛ༗͢Δ ೔ຊͷғޟAIΛݗҾ͠ɺ1100୆ ͷGPUΛ༻ֶ͍ͨशΛઃܭ ғޟɾকعAIͰੈքτοϓΫϥ εͷ࣮੷Λୡ੒ HEROZͷIPOΛϦʔυΤϯδ χΞͱͯ͠ݗҾ কعAIʮPonanzaʯͷ։ൃऀ ࢁຊҰ੒ ૑ۀऀCEO ాதେհ COO ੝ౡਖ਼ਓ CFO ࢁޱ༞ CTO ಙ஍༎ޛ CBO ϝυϨʔͷIPOΛࣥߦ໾һͱ͠ ͯݗҾ Google SuiteͷΤόϯδΣϦε τ HEROZ Google ϝυϨʔ ϞϧΨϯɾ ελϯϨʔ GEδϟύϯ HEROZ ࢈ۀٕज़ ૯߹ݚڀॴ όϯμΠ φϜί ϝσΟΞυΡ

Slide 33

Slide 33 text

ϝϯόʔͷόοΫάϥ΢ϯυ TURING COMPANY DECK | ORGANIZATION 33 AI Talent Software / Hardware Engineer Business / Corporate HEROZ ϦΫϧʔτ ౦ࣳ atama plus ౦ژେֶ ژ౎େֶ ۝भେֶ ஜ೾େֶ Kaggle Grandmaster αΠόʔΤʔδΣϯτ ΞΫηϯνϡΞ ηϯελΠϜδϟύϯ LINEϠϑʔ ೔మιϦϡʔγϣϯζ ཧԽֶݚڀॴ ΢ΥϯςουϦʔ Kaggle Master ϦΫϧʔτ αΠόʔΤʔδΣϯτ ιχʔ GROOVE X NTTίϛϡχέʔ γϣϯζ AWS Ξϓτϙου ຊాٕݚ޻ۀ LINEϠϑʔ λΠϛʔ ෋࢜௨ άʔάϧɾ Ϋϥ΢υδϟύϯ Preferred Networks IBM ౦ژେֶ άʔάϧ ϞϧΨϯɾελϯϨʔ ϝυϨʔ όϯμΠφϜί ΞΫηϯνϡΞ ΤΫα΢Οβʔζ ιϑτόϯΫ ύʔιϧΩϟϦΞ Luup ϦΫϧʔτ ΦϦϯύε ತ൛ҹ࡮ ϚωʔϑΥϫʔυ Speee READYFOR SmartHR

Slide 34

Slide 34 text

ಇ͘؀ڥɾ෱རްੜ TURING COMPANY DECK | WORK ENVIRONMENT 34 ΦϑΟεͱࣗ୐ͷͲͪΒͰ΋ۈ຿͕ՄೳͰ ͢ɻ ΫΠοΫʹίϛϡχέʔγϣϯͰ͖ΔͨΊ ͪΐͬͱͨٙ͠໰Λ͙͢ʹղফͰ͖·͢ɻ Ոఉͷ౎߹ʹ߹Θͤͯಇ͚Δ؀ڥ͕੔ͬͯ ͍·͢ɻ ϋΠϒϦουϫʔΫ ϑϨοΫελΠϜ੍౓ ࢈ҭٳɾհޢٳۀ੍౓ PC౳։ൃ؀ڥߏஙࢧԉ றं৔୅ิॿ੍౓ ΤϯδχΞ։ൃπʔϧࢧԉ ίΞλΠϜ͕11:00~15:00ͷϑϨοΫελ ΠϜ੍౓Ͱ͢ɻ  νʔϜϝϯόʔͱͷίϛϡχέʔγϣϯΛ ίΞλΠϜ಺Ͱߦ͏͜ͱʹΑͬͯɺͦΕҎ ֎ͷ࣌ؒ͸ࣗ਎ͷࡋྔͰॊೈʹಇ͍͍ͯ· ͢ɻ ࢈ٳɾҭٳɾհޢٳۀ੍౓͕͋Γɺऔಘ࣮ ੷΋ଟ͍Ͱ͢ɻ ೖࣾ1೥ະຬͷํͰ΋ɺ࠷௕2ϲ݄·ͰΛ্ ݶͱͯ͠ಛྫͰͷऔಘ࣮੷͕͋Γ·͢ɻ࠾ ༻໘઀࣌఺Ͱऔಘ༧ఆ͕ܾ·͍ͬͯΔํ͸ ॊೈʹରԠ͠·͢ͷͰɺ࠾༻ϓϩηεͷத Ͱ͓ਃ͠ग़͍ͩ͘͞ɻ ։ൃ෦ॴଐͷΤϯδχΞ͔ͭ30ࡀະຬͷํ Λର৅ͱͨ͠றं৔୅ิॿख౰Ͱ͢ɻ(্ ݶ5ສԁ/݄) एख։ൃऀ͕೔ৗతʹӡసͰ͖Δ؀ڥ࡞Γ Λࢧԉ͢Δ͜ͱͰɺࣗಈӡసٕज़։ൃͷߋ ͳΔ޲্Λ໨ࢦ੍͢౓Ͱ͢ɻ ChatGPT, Claude, Cursor, GitHub Copilot ͳͲͷ։ൃࢧԉπʔϧͷඅ༻Λձ͕ࣾෛ୲ ͢Δ੍౓͕͋Γ·͢ɻ ͜ΕΒͷ੍౓͸ΤϯδχΞɾϦαʔνϟʔ ͷ։ൃɾݚڀΛՃ଎͢Δ͜ͱΛ໨తͱͯ͠ ͍·͢ɻ ೖࣾ࣌ͷPCΛબ୒Ͱ͖Δ੍౓Λ༻ҙͯ͠ ͍·͢ɻ   ΦϑΟεͷσεΫτοϓ͸શͯ΢ϧτϥϫ ΠυσΟεϓϨΠΛ༻ҙ͠ɺࠓޙ΋։ൃʹ ઐ೦Ͱ͖Δ؀ڥࢧԉΛߦ͍·͢ɻ

Slide 35

Slide 35 text

ΤϯδχΞͱͯ͠ͷεΩϧΞοϓΛ࣮ݱ͢Δ؀ڥ TURING COMPANY DECK | WORK ENVIRONMENT 35 ʮEnd-to-EndʯࣗಈӡసγεςϜΛ։ൃ͍ͯ͠ Δձࣾ͸೔ຊͰνϡʔϦϯάͷΈʢถ͸Teslaɺ ӳ͸Wayveʣ ʮ׬શࣗಈӡసʯͱ͍͏ਓྨͷάϥϯυνϟϨ ϯδΛ࣮ݱ͢ΔͨΊʹ͸ɺҰਓͻͱΓʹେ୾ͳ νϟϨϯδ͕ٻΊΒΕ·͢ɻͦͷ෼ɺಘΒΕΔ ܦݧ͸ਓੜͰ͔͚͕͑ͷͳ͍ࢥ͍ग़ʹͳΔͱ৴ ͍ͯ͡·͢ɻ ·ͨɺࣗಈӡస × ੜ੒AIͷ։ൃ͸શੈքͰٸܹ ʹ஫໨͞Ε࢝ΊͨྖҬͰ͢ɻ͜ͷྖҬʹࠓඈͼ ࠐΉ͜ͱͰಘΒΕΔ஌ݟ΍ܦݧ͸ࠓޙͷਓੜΛ ๛͔ʹͯ͘͠ΕΔͰ͠ΐ͏ɻ ੈքΫϥεͷେ͖ͳ໨ඪ τοϓΤϯδχΞͱେن໛ͳܭࢉࢿݯ ւ֎ग़ுɾΧϯϑΝϨϯεొஃ ■ AI / ιϑτ΢ΣΞʹڧ͍ܦӦνʔϜ ■ ೔ຊ΍ੈքΛ୅ද͢ΔIT/Webاۀ  ɹग़਎ऀ ■ ೔ຊΛ୅ද͢ΔϝʔΧʔग़਎ऀ ͳͲ֤اۀͷτοϓΤϯδχΞ͕ू·Δଟ༷ੑ ͷ͋ΔΤϯδχΞ૊৫Ͱ͢ɻ ·ͨɺܦࡁ࢈ۀলͷGENIACʹ࠾୒͞ΕͨΓɺ ࣗࣾGPUΫϥελʢH100 96جʣΛߏஙͨ͠Γ ͱɺେن໛ͳܭࢉࢿݯΛ֬อ͠׆༻͢Δମ੍Λ ࡞͍ͬͯ·͢ɻ1ਓ౰ͨΓͷΤϯδχΞ͕׆༻Ͱ ͖Δܭࢉࢿݯྔ͸ࠃ಺τοϓΫϥεͰ͢ɻ ࠷৽ͷࣗಈӡసγεςϜͷࢹ࡯΍CVPR౳ͷ τοϓΧϯϑΝϨϯε΁ࢀՃ͢ΔͨΊͷւ֎ग़ ு͕ՄೳͰ͢ɻ৽ଔࣾһͷଟ͘΋ւ֎ग़ுΛܦ ݧ͓ͯ͠Γɺࠃ಺ʹཹ·ΒͣੈքΛҙࣝͨ͠։ ൃ΍ݚڀʹ஫ྗͰ͖·͢ɻ ࠃ಺ͷΧϯϑΝϨϯε΍AWS Summit౳ʹ΋ੵ ۃతʹࢀՃ͓ͯ͠Γɺొஃ͢Δػձʹܙ·ΕΔ ؀ڥͰ͢ɻ

Slide 36

Slide 36 text

36 ࠾ ༻ TURING COMPANY DECK | CAREERS CAREERS

Slide 37

Slide 37 text

ߦಈࢦ਑ TURING COMPANY DECK | VALUES 37 ώϡʔϦεςΟοΫͳௐ੔Ͱ͸ͳ͘ɺܭࢉྔʹΑΔεέʔϦϯά͕ͦ͜௕ظతʹ ͸উར͠·͢ɻখखઌͷ޻෉ʹಀ͛ͣɺ൚༻తͰຊ࣭తͳղܾࡦΛબͼ·͠ΐ ͏ɻ※͜ͷݴ༿͸ɺڧԽֶशͷ෕ͱݺ͹Ε2024೥ʹνϡʔϦϯά৆΋ड৆ͨ͠ϦονɾαοτϯʢRich Suttonʣࢯͷ༗ ໊ͳΤοηΠʮThe Bitter LessonʯʹͳΒ͍ͬͯ·͢ɻ The Bitter Lesson όϦϡʔ Less is More ϚγϡϚϩΛͯ͞͠ΈΑ͏ νʔϜ͸খ͘͞ɺૉૣ͘ɺมԽ͢Δ ͦΕͰ΋ɺલʹਐΉ ίʔυ΋Ϟδϡʔϧ΋ձٞ΋ɺಉ͡੒ՌͳΒʮগͳ͍ʯ͜ͱ͕ਖ਼ٛͰ͢ɻৗʹʮ࡟ Εͳ͍͔ʁʯͱ໰͍͔͚ɺෳࡶ͞Λഉআͯ͠ຊ࣭తͳՁ஋͚ͩʹूத͠·͠ΐ ͏ɻ ׬ᘳͳܭըΑΓ΋ɺ·ͣ͸खΛಈ͔ͯ͠ܗʹͯ͠ΈΔ͜ͱ͕େ੾Ͱ͢ɻ໎͍΍ઌ Ԇ͹͠ʹΑΔίετΛܰࢹͤͣɺૣظͷ࣮ߦͱվળΛ܁Γฦ͠·͠ΐ͏ɻ ※νϡʔϦϯάͰ͸ʮϚγϡϚϩɾνϟϨϯδʯͱ͍͏༗໊ͳνʔϜϏϧσΟϯάͷήʔϜΛ࣮ࢪ͍ͯ͠·͢ɻ νʔϜ͸ৗʹ࠷খԽ͠ɺϦʔμʔΛదਖ਼ʹԠͯ͡ੵۃతʹൈṅɾަ୅͠·͢ɻܹ ͍͠มԽͷதͰ΋ޓ͍ʹϦεϖΫτΛ࣋ͪɺର࿩Λઈ΍ͣ͞࿈ܞ͠·͠ΐ͏ɻ େ͖ͳมԽ΍ਐԽʹ͸ɺ͠͹͠͹ຎࡲ΍௧Έ͕൐͍·͢ɻࠔ೉ͳҙࢥܾఆ΍ݫ͠ ͍ݱ࣮͔Βಀ͛ͣɺ௧ΈΛ੒௕ͷྐͱͯ͠ड͚ೖΕɺͦΕͰ΋લʹਐΈ·͠ΐ ͏ɻ

Slide 38

Slide 38 text

■ ࠾༻Ϛωʔδϟʔ ■ ࠾༻୲౰ ■ ϑΝΠφϯεϚωʔδϟʔ ■ ܦཧϚωʔδϟʔ ■ MLΤϯδχΞ ■ RLΤϯδχΞ ■ MLOps ■ σʔλΤϯδχΞɺΞφϦςΟΫεΤϯδχΞ ■ ιϑτ΢ΣΞΤϯδχΞ(WebɺLinuxɺSRE) ■ ιϑτ΢ΣΞΤϯδχΞ(OSɾ૊ΈࠐΈɺϛυϧ΢ΣΞɺϑΝʔϜ ΢ΣΞɺϩϘςΟΫε) ■ ϑϧελοΫΤϯδχΞ(Web) ■ ੍ޚΤϯδχΞ ■ GPUΫϥελ / HPCΫϥελΤϯδχΞ ■ ࣗಈӡసAI։ൃυϥΠόʔ ■ ৽ଔΤϯδχΞ ืू৬छ TURING COMPANY DECK | CAREERS 38 Ԡืϖʔδ͸ͪ͜Β

Slide 39

Slide 39 text

TURING COMPANY DECK | CAREERS 39 ॊೈͳબߟϑϩʔ ࠨʹهࡌ͍ͯ͠Δબߟϑϩʔ͸͋͘·Ͱ΋جຊతͳܗࣜͰ͢ɻ Έͳ͞Μͷ͝ঢ়گʹԠͯ͡ɺ໘ஊɾΦϑΟεݟֶɾOpen OfficeͳͲͷௐ੔͕ՄೳͰ͢ɻ ΋ͪΖΜଞࣾબߟঢ়گʹԠͯ͡౰ࣾͷબߟϓϩηεΛૣΊ͍ͨ౳ͱ͍͏͝૬ஊ΋ՄೳͰ͢ɻ ※બߟظؒͷ໨҆͸ɺ͓Αͦ2ʙ3िؒͰ͢ɻ STAR໘઀(ߦಈ໘઀) νϡʔϦϯάͰ͸ɺաڈͷߦಈΛ۷ΓԼ͛Δ໘઀ख๏ͱͯ͠ʮSTARʯΛ׆༻͍ͯ͠·͢ɻ ͲͷΑ͏ͳঢ়گ(Situation)ͰɺͲͷΑ͏ͳ՝୊(Task)ʹରͯ͠ɺ ͲͷΑ͏ͳߦಈ(Action)Λى͜͠ɺͲͷΑ͏ͳ݁Ռ(Result)Λग़ͨ͠ͷ͔ɻ ͜ΕΒʹؔ͢Δ࣭໰Λ͢Δ͜ͱͰɺօ͞Μͷܦݧ΍εΩϧΛਖ਼͘͠ཧղ͍ͨ͠ͱߟ͍͑ͯ·͢ɻ ମݧೖࣾ ମݧೖࣾͰ͸ɺSlack΍NotionͳͲήετݖݶͱͯࣾ͠಺৘ใʹΞΫηεͰ͖ΔΑ͏ʹ͠ɺ഑ଐ༧ఆ νʔϜͷࣾһͱ൒೔ or 1೔ؒҰॹʹಇ͍͍ͯͨͩ͘ϓϩηεͰ͢ɻ νϡʔϦϯάͷΧϧνϟʔΛମײ͍͖ͨͩɺ͓ޓ͍ʹϛεϚον͕ͳ͍͔Λ֬ೝ͢ΔͨΊͷ৔ʹͳΓ ·͢ɻ Recruiting Process ※Πϯλʔϯͷ৔߹͸࠷௕2೔ؒ ※ٕज़ࢼݧ͸഑ଐνʔϜʹԠ࣮ͯ͡ ࢪ͢Δ͔Ͳ͏͔ܾΊ·͢

Slide 40

Slide 40 text

40 ■ ৬ྺ֓ཁ  ɹ- ͜Ε·Ͱͷܦྺ΍࣮຿ܦݧɺಘҙྖҬɾڵຯྖҬʹ͍ͭͯ200~300จ ࣈఔ౓ͷ·ͱΊ ■ εΩϧ  ɹ- ͜Ε·Ͱ࢖͖ͬͯͨ։ൃݴޠ΍ܞΘ͖ͬͯͨ։ൃྖҬɺ͓Αͼͦͷश ख़౓ʹ͍ͭͯ  ɹ- ͜Ε·Ͱܦݧ͖ͯͨ͠։ൃ؀ڥ΍໾ׂɺ੹೚ൣғʹ͍ͭͯ ■ ৬ྺ֓ཁ  ɹ- ॴଐاۀຖͷ৬຿಺༰ɿಛʹͲ͏͍ͬͨ։ൃ؀ڥɾνʔϜମ੍ͰɺͲ Μͳ໾ׂΛ୲͍ԿΛ։ൃͨ͠ͷ͔  ɹ- ։ൃ੒Ռ෺͕෼͔Δࢿྉ΍URLʢαʔϏεαΠτ΍GitHubɺςοΫϒ ϩάͳͲʣ Title Title Title | CONTENTS νϡʔϦϯά͕ॻྨબߟ࣌ʹ஫໨͍ͯ͠ΔϙΠϯτ ৬຿ܦྺॻʹهࡌͯ͠΄͍͠৘ใɾମݧೖࣾʹ͍ͭͯ ମݧೖࣾʹ͍ͭͯ ■ ࣮ࢪ໨తɾ֓ཁ  ମݧೖࣾ͸૒ํͷϛεϚονΛ๷͙͜ͱΛ໨తͱ͍ͯ͠·͢ɻ׬શࣗಈӡ సͱ͍͏૖େͳϛογϣϯʹ௅Ήʹ͸ɺԠืऀ༷΍ͦͷ͝Ո଒ͷօ༷͕ೲ ಘ͢ΔܗͰೖࣾҙࢥΛܾΊ͍ͯͨͩ͘͜ͱ͕ඇৗʹॏཁͩͱߟ͍͑ͯ· ͢ɻͦͷͨΊɺࣄલʹNDAΛక্݁ͨ͠ͰɺSlack΍NotionͳͲମݧೖࣾ΁ ͓ӽ͍͍ͨͩͨ͠ํͷΈʹެ։Ͱ͖Δൣғͷࣾ಺৘ใΛ͝ཡ͍ͨͩ͘Α͏ ʹઃܭ͍ͯ͠·͢ɻ ·ͨɺ౰೔͸഑ଐ૝ఆνʔϜ͕ࣄલʹ༻ҙͨ͠՝୊ʹऔΓ૊ΜͰ͍ͨͩ ͖ɺऔΓ૊Έ಺༰Λൃද͍ͨͩ͘͜ͱΛ༧ఆ͍ͯ͠·͢ɻʢ՝୊ʹ͍ͭͯ͸ ౰೔ޕલதʹڞ༗͠·͢ʣ ͦͷଞɺνϡʔϦϯάͷ࣮ࡍͷ։ൃ؀ڥ΍૊৫จԽͳͲΛഽ਎Ͱײ͡औͬ ͯ΋Β͏͜ͱͰɺస৬ઌͱͯ͠ద͍ͯ͠Δ͔Λ֬ೝ͢Δػձͱͯ͠΋ଊ͑ͯ ͍ͨͩ͘͜ͱΛظ଴͍ͯ͠·͢ɻ

Slide 41

Slide 41 text

TURING COMPANY DECK | CAREERS 41 ετοΫΦϓγϣϯ ετοΫΦϓγϣϯͱ͸ɺैۀһʹରͯ͠اۀͷגࣜΛҰఆͷՁ֨Ͱߪೖ͢ΔݖརΛ෇༩͢Δ੍౓Ͱ ͢ɻ ͜ͷ࢓૊ΈͷҰྫͱͯ͠ɺ೔ຊΛ୅ද͢ΔελʔτΞοϓͷҰͭɺϝϧΧϦ͕͋Γɺ30໊Ҏ্ͷैۀ һ͕6ԯԁҎ্ͷࢿ࢈Λ֫ಘͨ͠ͱݴΘΕ͍ͯ·͢ɻνϡʔϦϯάͷ૑ۀ΋ɺ୅දͷࢁຊ͕લ৬Ͱ͋Δ HEROZͷ্৔ɾετοΫΦϓγϣϯ׆༻ʹΑͬͯಘΒΕͨࢿ࢈͕ݪࢿͱͳ͍ͬͯ·͢ɻ γϦʔζAͰͷେܕௐୡ΋׬ྃ͠ɺετοΫΦϓγϣϯͷ഑෍͸͜Ε͔Βߦ͍·͢ɻνϡʔϦϯάͰ͸ ՄೳͳݶΓͨ͘͞ΜͷϝϯόʔʹετοΫΦϓγϣϯΛ഑Γɺ࣍ੈ୅࢈ۀͷҭ੒΍৽ͨͳΤίϊϛΫ ε૑ग़ʹ௅ઓ͍͖ͯ͠·͢ɻΘΕΘΕ͕ఆΊΔ৚݅͸ԼهͰݕ౼͍ͯ͠·͢ɻ ■ νϡʔϦϯά্͕৔͢Δ͜ͱ ■ SO෇༩࣌ʹਖ਼ࣾһͰ͋Δ͜ͱ ■ ෇༩͢ΔSO͸ݪଇ੫੍ద֨SO ■ SO෇༩͔Β5೥ؒܦաͰ100%ͷݖརߦ࢖͕Մೳʢຖ೥20%෼ͷݖར֬ఆʣ ■ ୀ৬࣌ʹͦΕ·Ͱʹݖར֬ఆͨ͠SO͸࣋ͪग़͠Մೳ ·ͨɺৄࡉͳظ଴஋΍ετοΫΦϓγϣϯ੍౓ͷߏஙഎܠ͸υΩϡϝϯτͰ·ͱΊΒΕɺࣾһʹެ։͞Ε͍ͯ·͢ɻ ※ࣾһ਺૿ՃɾࣄۀϑΣʔζมԽʹΑΓετοΫΦϓγϣϯͷ഑෍͕ͳ͘ͳΔՄೳੑ͕͋Γ·͢ɻ ※STOCK JOURNALʮ“͋·Γ࣮ײΛ࣋ͯͳ͔ͬͨSO”͕ɺ׬શࣗಈӡసͱ͍͏Turingͷ௅ઓʹͭͳ͕ͬͨ࿩ʯ ΋ซͤͯ͝ཡ͍ͩ͘͞ɻ هࣄ͸ͪ͜Β