A Journey as Staff Engineer at SmartNews! 〜一年間の経験から語る、ICキャリアの今とこれから〜

Scrum Fest Mikawa 2025 の発表資料です。

Transcript

1. ʙҰ೥ؒͷܦݧ͔ΒޠΔɺ*$ΩϟϦΞͷࠓͱ͜Ε͔Βʙ September 15, 2025 A Journey as Staff Engineer at SmartNews!
2. Who am I? / ͓·ͩΕ Ikuo Suyama / ಃࢁ
   ҭஉ • Staff Engineer • Ads Backend Expert • Nov. 2020~ SmartNews, Inc. • Interest: Fishing, Camping, Gunpla, Anime
3. Who am I? / ͓·ͩΕ Ikuo Suyama / ಃࢁ
   ҭஉ • Ads Backend Expert • Nov. 2020~ SmartNews, Inc. • Interest: Fishing, Camping, Gunpla, Anime •Staff Engineer
4. 4 ࠓ೔͸ Staff Engineer … ͷ࿩Λ͠·͢ʂ
5. 5 ՝୊ҙࣝͱϞνϕʔγϣϯ Over the past few years we've seen a flurry of books unlocking the engineering manager career path, like: The Manager's Path by Camille Fournier, The Making of a Manager by Julie Zhuo, and my own, An Elegant Puzzle. The management career isn't an easy one, but increasingly there is a map available. …The transition into Staff Engineer, and its further evolutions like Principal Engineer, remains particularly challenging and undocumented. ͜͜਺೥ɺΤϯδχΞϦϯάϚωʔδϟʔͷΩϟϦΞύεΛղ͖໌͔͢ຊ͕࣍ʑͱग़൛͞Ε͍ͯ·͢ɻྫ ͑͹ɺ… ϚωδϝϯτͷΩϟϦΞ͸ܾͯ͠؆୯ͳಓͰ͸͋Γ·ͤΜ͕ɺঃʑʹͦͷʮ஍ਤʯ͕੔͖͍ͬͯͯ·͢ɻ …ҰํͰɺStaff Engineer ΁ͷҠߦɺ͞ΒʹͦͷਐԽܗͰ͋Δ Principal Engineer ΁ͷಓͷΓ͸ɺґવͱ ͯ͠ಛʹ೉͘͠ɺे෼ʹه࿥͞Ε͍ͯͳ͍ঢ়گͰ͢ɻ — Staff Engineer Leadership beyond the management track ᶃ ଘࡏΛ஌ͬͯ΋Β͍͍ͨʂ ᶄ ΤϯδχΞΩϟϦΞύεͷબ୒ࢶ͕૿͑Δͱخ͍͠ʂ ᶅ ϙδγϣϯ͕૿͑Δͱخ͍͠ʂ
6. 6 ͦ͜ʹݱΕͨૉ੖Β͍͠ॻ੶ʂ “ελοϑΤϯδχΞ ϚωδϝϯτΛ௒͑ΔϦʔμʔγοϓ” 2023/05 “ελοϑΤϯδχΞͷಓ ―༏Εٕͨज़ઐ໳৬ʹͳΔͨΊͷΨΠυ” 2024/08
7. 7 ͦ͜ʹݱΕͨૉ੖Β͍͠ॻ੶ʂ “ελοϑΤϯδχΞ ϚωδϝϯτΛ௒͑ΔϦʔμʔγοϓ” 2023/05 “ελοϑΤϯδχΞͷಓ ―༏Εٕͨज़ઐ໳৬ʹͳΔͨΊͷΨΠυ” 2024/08 Staffʹ ͳΔ·Ͱͷ࿩ Staffʹ ͳ͔ͬͯΒͷ࿩
8. 8 ͦ͜ʹݱΕͨૉ੖Β͍͠ॻ੶ʂ “ελοϑΤϯδχΞ ϚωδϝϯτΛ௒͑ΔϦʔμʔγοϓ” 2023/05 “ελοϑΤϯδχΞͷಓ ―༏Εٕͨज़ઐ໳৬ʹͳΔͨΊͷΨΠυ” 2024/08 Staffʹ ͳΔ·Ͱͷ࿩ Staffʹ ͳ͔ͬͯΒͷ࿩ ΄ͱΜͲ Blog staffeng.com ͰಡΊ·͢(ӳޠ)
9. 1.Staff Engineer ͱ͸ʁ 2.Ͳ͏΍ͬͯͳΔʁ 3.ͲΜͳ࢓ࣄΛ͍ͯ͠Δʁ 4.͜Ε͔Βͷੜଘઓུʁ Agenda
10. 01 Staff Engineer … ͱ͸ʁ
11. 11 1-1. Staff Engineer ͷҙຯͱ༝དྷ Staff Engineerͱ͸ʁ Individual Contributor (ΠϯσΟϏδϡΞϧɾίϯτϦϏϡʔλʔIC) …ͷ্Ґ৬
12. 12 1-1. Staff Engineer ͷҙຯͱ༝དྷ Individual Contributor ͱ͸ʁ Manager(؅ཧ৬) Ͱ͸ͳ͍ ٕज़ઐ໳৬ͱͯ͠ͷΩϟϦΞ(ύε)
13. 13 1-1. Staff Engineer ͷҙຯͱ༝དྷ Staff Officer: ࢀ๳কߍ ໊લͷ༝དྷ “an army officer who helps the officer in charge to plan military activities” ࢦش׭ͷԼͰ܉ࣄ׆ಈͷܭըΛࢧԉ͢Δকߍʢࢀ๳কߍʣɻ — Cambridge Dictionary ࢘ྩ׭Λܭըɾ෼ੳͰิࠤ͢Δٕज़ࢀ๳ సͯ͡ɺݱ୅ιϑτ΢ΣΞ૊৫ʹ͓͍ͯ
14. 14 1-2. ཱͪҐஔͱظ଴͞ΕΔ੹೚ ΠϝʔδɿνʔϜΑΓେ͖ͳ૊৫ΛݟΔ ”TechLead” — Staff Engineer: Leadership beyond the management track / Introduction Α͋͘Δ IC্ڃ৬(Staff+)ͷ໊લ
15. 15 1-2. ཱͪҐஔͱظ଴͞ΕΔ੹೚ — Staff Engineer: Leadership beyond the management track / Introduction Team Group Division Department (܎௕?) (՝௕?) (෦௕?) (ہ௕?) ͨͿΜ͍͍ͩͨ ͜͏͍͏ײ͡ͷ෼ঠ ΠϝʔδɿνʔϜΑΓେ͖ͳ૊৫ΛݟΔ ”TechLead”
16. 16 1-2. ཱͪҐஔͱظ଴͞ΕΔ੹೚ — Staff Engineer: Leadership beyond the management track / Introduction Team Group Division Department Senior ɹ Staff Senior ɹ Staff Group Head Manager Division Head(VPoE) CTO (܎௕?) (՝௕?) (෦௕?) (ہ௕?) SmartNews ͷ͹͍͋ ΠϝʔδɿνʔϜΑΓେ͖ͳ૊৫ΛݟΔ ”TechLead”
17. 17 1-2. ཱͪҐஔͱظ଴͞ΕΔ੹೚ — Staff Engineer: Leadership beyond the management track / Introduction Team Group Division Department Senior ɹ Staff Senior ɹ Staff Group Head Manager Division Head(VPoE) CTO (܎௕?) (՝௕?) (෦௕?) (ہ௕?) Ikuo ΠϚίίʂ ΠϝʔδɿνʔϜΑΓେ͖ͳ૊৫ΛݟΔ ”TechLead”
18. 18 1-2. ཱͪҐஔͱظ଴͞ΕΔ੹೚ Manager ͱͷҧ͍… ؍఺ Staff / Staff-plusʢICʣ ManagerʢEM/Director ͳͲʣ ओͨΔ੹຿ ٕज़ํ਑ͷઃఆɾฤूɺϝϯλϦϯά/εϙϯα ʔγοϓɺ૊৫ͷҙࢥܾఆ΁ٕज़จ຺Λ஫ೖɻ ਓͱ࢓ࣄͷϚωδϝϯτʢௐ੔ɾӡ༻ʣɻ ͠͹͠͹ௐ੔/ࣄ຿ͷߑਫʹࡽ͞Ε͕ͪɻ ݖݶͷݯઘ “ҕ೚ʢproxiedʣ͞Εͨݖݶ”ʹґଘɻ௨ৗ͸Ϛ ωʔδϟͳͲͷϦʔμʔ͔Βͷҕ೚ͰӨڹྗΛൃ ش͢Δɻ ૊৫্ͷਖ਼ࣜͳݖݶʢ্ه “ҕ೚ݩ” ଆʣɻ Staff ͱ࿈ܞͭͭ͠ྖҬͷ࠷ऴ੹೚Λෛ͏ɻ είʔϓ νʔϜ/ෳ਺ྖҬʹԣஅٕͯ͠ज़ํ޲ΛݗҾʢಛ ఆྖҬͷϦʔυɺ·ͨ͸޿ൣғͷ੔߹औΓʣɻ νʔϜ/૊৫ͷӡӦͱ༏ઌॱҐ෇͚ɻඞཁʹ Ԡͯ͡ ٕज़඼࣭ͷ੹຿Λ Staff ʹ“ৡΔ”/೚ͤ Δɻ Өڹͷग़͠ํ ࣗ෼Ͱ͸ͳ͘“पғΛ௨ͯ͡”੒ՌΛ૿෯ʢϝϯλ ϦϯάΑΓ΋εϙϯαʔγοϓΛް͘ɺٕज़ࢹ఺ ͰҙࢥܾఆΛޙԡ͠ʣɻ ໾্ׂͷ؅ཧɾௐ੔Ͱ࣮ߦྗΛੜΉʢͨͩ͠ “؅ཧ࢓ࣄͷ߃ৗతͳؙ౤͛” ͸ආ͚Δʣɻ ਓͷ؅ཧ ICʢIndividual ContributorʣͰ͋Γɺ௚઀ͷϐʔ ϓϧϚωδϝϯτ͸໾ׂ֎ɻ ධՁɾฤ੒ɾ༏ઌ౓ௐ੔ͳͲͷϐʔϓϧ/Φϖ Ϩʔγϣϯ؅ཧ͕Ұ࣍੹຿ɻ Staff Engineer What do Staff engineers actually do? / Frequently Asked Questions
19. 19 1-2. ཱͪҐஔͱظ଴͞ΕΔ੹೚ Manager ͱͷ Overlap Engineering Management Triangle “Management” Λ ΍Βͳ͍ͳ͍Θ͚Ͱ͸ͳ͍
20. 20 1-2. ཱͪҐஔͱظ଴͞ΕΔ੹೚ Manager ͱͷ Overlap Engineering Management Triangle ओઓ৔ Impactग़͢ͷʹ ඞਢ ΍Δ͚Ͳ ϝΠϯͰ͸ͳ͍ ਓࣄɾධՁ͸ ੹຿֎ ͜ͷ΁Μ͕ ओ੹຿ Managerͱ ڠۀ
21. 21 Staff Engineer — Staff archetypes 3. Solverʢιϧόʔʣ • ໾ׂɿඇৗʹෳࡶͳ໰୊ʹਂ͘౿ΈࠐΈɺղܾͷಓےΛݟ͚ͭΔεϖγϟ Ϧετɻ • ಛ௃ɿ૊৫͕༏ઌ͢Δະղܾͷ՝୊ʹऔΓ૊Έɺ໰୊͕ऩଋ͢Ε͹ผͷ ʮϗοτεϙοτʯʹҠಈ͢ΔελΠϧɻνʔϜ୯ҐΑΓݸਓ΁ͷґଘܕ Ͱ͋Δ͜ͱ͕ଟ͍ ɻ 2. ArchitectʢΞʔΩςΫτʣ • ໾ׂɿ૊৫ͷॏཁͳٕज़ྖҬʹରͯ͠ɺํ޲ੑɾ඼࣭ɾΞϓϩʔνΛ ୲อɻ • ಛ௃ɿϏδωεχʔζɾϢʔβʔཁ݅ɾٕज़੍໿Λਂ͘ཧղͨ͠ɺෳ ਺νʔϜΛӽٕ͑ͨज़ઓུͷਪਐऀɻେن໛ɾෳࡶͳ؀ڥԼͰ׆༂ɻ 4. Right Handʢӈ࿹ʣ • ໾ׂɿΤάθΫςΟϒͷิࠤ໾ͱͯ͠ɺ૊৫಺֎ͷෳࡶͳ໰୊ʹରԠ͠ɺ ϦʔμʔͷӨڹྗΛ֦ு͢Δɻ • ಛ௃ɿٕज़͚ͩͰ͸ͳ͘ɺϏδωεɺจԽɺਓͷ໰୊·Ͱ෯޿ؔ͘༩ɻ༏ ઌ౓ͷߴ͍՝୊ʹରॲ͠ɺ௚઀తͳϚωδϝϯτΛ࣋ͨͣʹڧ͍ӨڹྗΛ ൃشɻ 1-3. Staff Engineer ͷ 4ͭͷ "ΞʔΩλΠϓ" 1. Tech LeadʢςοΫϦʔυʣ • ໾ׂɿಛఆνʔϜͷΞϓϩʔνͱ࣮ߦΛಋ͘ɻϚωʔδϟʔͱ ࿈ܞ͠ͳ͕ΒɺνʔϜͷٕज़తϏδϣϯ΍໨ඪΛڞ༗ɾ਱ߦɻ • ಛ௃ɿෳࡶͳλεΫΛείʔϓͯ͠ௐ੔͠ɺνʔϜͷো֐Λഉ আɻίʔσΟϯάओಋ͔Βϝϯόʔ΁ͷҕৡ͕த৺ʹɻ
22. 22 1-3. Staff Engineer ͷ 4ͭͷ "ΞʔΩλΠϓ" όΠφϦͰ͸ͳ͍ Solver 導 築 解 補 Tech Lead νʔϜΛಋ͖ɺ ํ޲ੑΛࣔ͢ Architect ٕज़Λઃܭ͠ɺ ࢓૊ΈΛܗͮ͘Δ Right Hand ܦӦ૚Λิࠤ͠ɺ શମΛࢧ͑Δ ෳࡶͳ໰୊Λղ͖΄͙͢
23. 23 1-3. Staff Engineer ͷ 4ͭͷ "ΞʔΩλΠϓ" 導 築 解 補 όΠφϦͰ͸ͳ͍ Tech Lead νʔϜΛಋ͖ɺ ํ޲ੑΛࣔ͢ Architect ٕज़Λઃܭ͠ɺ ࢓૊ΈΛܗͮ͘Δ Right Hand ܦӦ૚Λิࠤ͠ɺ શମΛࢧ͑Δ Solver ෳࡶͳ໰୊Λղ͖΄͙͢ Ikuo ྫɿIkuo ݩͷνʔϜʹ࣠଍Λஔ͖ͭͭɺ ΑΓ޿͍Division/Department ʹӨڹͷ ͋Δ࢓૊Έͮ͘ΓΛ͍ͯ͠Δ • Πϯϑϥίετ෼ੳɾ࡟ݮͱΨόφϯε • ΦϒβʔόϏϦςΟͷ޲্ Etc…
24. 24 1-3. Staff Engineer ͷ 4ͭͷ "ΞʔΩλΠϓ" 導 築 解 補 όΠφϦͰ͸ͳ͍ Tech Lead νʔϜΛಋ͖ɺ ํ޲ੑΛࣔ͢ Architect ٕज़Λઃܭ͠ɺ ࢓૊ΈΛܗͮ͘Δ Right Hand ܦӦ૚Λิࠤ͠ɺ શମΛࢧ͑Δ Solver ෳࡶͳ໰୊Λղ͖΄͙͢ Ikuo Senior Staff A Senior Staff B Staff C ਂ͍υϝΠϯ஌ࣝͱ ໰୊ղܾೳྗ νʔϜʹਂ͍ಎ࡯
25. 25 1-3. Staff Engineer ͷ 4ͭͷ "ΞʔΩλΠϓ" όΠφϦͰ͸ͳ͍ • ӨڹྗΛͲ͜ͰɺͲͷΑ͏ʹൃش͢Δ͔ͱ͍͏࿩ • ྡΓ͋ͬͨΞʔΩλΠϓ͸ൃش͠΍͍͢ • Ͳ͏બͿ͔ʁ • ࢿ࣭΋͋Δ͕ɺܦݧ΋Өڹେ • ͍·ैࣄ͍ͯ͠ΔϓϩδΣΫτʹ΋ΑΔ͕… ICΩϟϦΞΛߟ͑Δ͏͑Ͱҙ͓ࣝͯ͘͠ͱྑ͍ ʢ͔΋ 導 築 解 補
26. 26 1-3. Staff Engineer ͷ 4ͭͷ "ΞʔΩλΠϓ" IkuoྲྀɿΞʔΩλΠϓ “࣌ݟࣜ” — ࣌ؒͷ࢖͍ํ͸ʁ Staff Engineer — Staff archetypes Category / Archetype A. ࣮૷ɾݸਓ࡞ۀ B. ઃܭɾϨϏϡʔ C. ࣮ߦਪਐɾௐ੔ D. ૊৫ɾར֐ௐ੔ Tech Lead 20% 30% 35% 15% Architect 10% 50% 20% 20% Solver 55% 25% 10% 10% Right Hand 5% 10% 35% 50% • ࣮૷ɾݸਓ࡞ۀ ➡ ղ: Solver • ઃܭɾϨϏϡʔ ➡ ங: Architect • ૊৫ɾར֐ௐ੔ ➡ ิ: Right Hand • όϥϯε ➡ ಋ: Tech Lead ͬ͘͟ΓҰ൪ଟ͍࣌ؒͷ࢖͍ํ͕… Summarized by ChatGPT 5-thinking
27. Ͳ͏΍ͬͯ 02 …Staff Engineer ʹͳΔʁ
28. 28 ॏཁɿ ICͱͯ͠ͷΩϟϦΞύεɺϥμʔ͕ଘࡏ͢Δ ձࣾʹॴଐ͢Δ ࠷ॳͷҰา ϥμʔΛθϩ͔Β “࡞Δ” ͷ͸ICʹ͸ݫ͍͔͠΋… ͨͩ͠৬ҐΛ্͛ΔͨΊ͚ͩͷస৬͸جຊ͓͢͢Ί͠ͳ͍ ※ݸਓͷײ૝Ͱ͢ɻ 2-1. جຊઓུ
29. 29 2-1. جຊઓུ Promotion Packets — ঢਐܭըɾূ੻υΩϡϝϯτ • ࣗ෼ͷ࢓ࣄΛΞϐʔϧ͢ΔυΩϡϝϯτΛ४උ͢Δ • ͋ͳͨͷʢStaffʣϓϩδΣΫτ͸ʁ • ͲͷΑ͏ʹձࣾΛվળͨ͠ʁ • ͲΜͳΠϯύΫτʢ਺ࣈʣ͕͋ͬͨʁ • ୭Λࢦಋͨ͠ʁ • ࣗ෼ʹ଍Γͳ͍ٕೳ΍ߦಈΪϟοϓ΁ͷରԠ͸ʁ Staff Engineer — Promotion packets • “ࠓ” ϓϩϞʔγϣϯʹڵຯ͕ͳͯ͘΋ɺఆظతʹ࡞ͬͯߋ৽͢Δͷ͕͓͢͢Ί • ࠓͷϚωδϟʔΛר͖ࠐΈɺҰॹʹ࡞੒͢Δͷ͕͓͢͢Ί • ϚωδϟʔʹϓϩϞʔγϣϯ΁ͷҙཉΛ఻͑ɺڠྗΛڼ͙ ݏΒ͘͠ฉ͑͜Δ͔΋͠Εͳ͍͕ɺ ΩϟϦΞΞοϓʹ͓͍ͯࣗݾΞϐʔϧ͸௒ॏཁ
30. 30 2-2. ࣮ྫɿIkuoͷ৔߹ ̏೥લɿνʔϜ Tech Lead / Managerͱͷ 1on1 … ͤ΍͔ͯ Ikuo… Ikuo ౰࣌ͷManager ࣗ෼ͷϓϩϞʔγϣϯʹ͸ڵຯͳ͍͔Βɺ νʔϜͷ͍͋ͭͱ͍͋ͭΛ͸΍͘ Senior ʹ͠Α͏
31. 31 • ʮࣗ෼ͷ Promotion ʹ͸ڵຯ͕ͳ͍ʯ࣮ࡍʹͳ͔ͬͨ • લఏͱͯ͠ɺICͱͯ͠ੜ͖ΔͱܾΊ͍ͯͨ (2019 Blog:σϕϩούʔͱͯ͠ੜ͖͍ͯ͘) • څྉ͸े෼΋ΒͬͯΔ͠… ࠓͷ࢓ࣄ͸ָ͍͠͠… • ݁ՌΛग़ͯ͠Ε͹ͦͷ͏ͪධՁ͞ΕΔ͸ͣɺͦ͏ߟ͍͑ͯͨ • Senior ʹ্͕ͬͨͱ͖͕ͦ͏ͩͬͨ • Կ͕ى͔ͬͨ͜ʁ̏೥ؒ Senior Ͱ Stuck • Staff ΁ͷঢਐ͸উखʹ/ࣗಈͰ͸ਐ·ͳ͍ ̏೥લɿνʔϜ Tech Lead / Managerͱͷձ࿩… ঢਐ͚͕ͩશͯͰ͸ͳ͍͕ɺ (ಛʹStaff+΁ͷ)ΩϟϦΞΞοϓʹ͓͍ͯࣗݾΞϐʔϧ͸௒ॏཁ 2-2. ࣮ྫɿIkuoͷ৔߹
32. 32 2-2. ࣮ྫɿIkuoͷ৔߹ 2೥લɿॻ੶ “Staff Engineer” ͱͷग़ձ͍ ͜Μͳಓ͕͋ͬͨΜ͔… ͜Ε΍ɺ ͜Εͦ͜Զͷੜ͖Δಓ…ʂʂ ͨͬ͢; ͑Μ͡ʹ͋ ॳΊͯICͱͯ͠ͷ۩ମతͳΩϟϦΞύεΛ͸͖ͬΓͱೝࣝͨ͠ʂ Ikuo
33. 33 2-2. ࣮ྫɿIkuoͷ৔߹ There’s a will, There’s a way! Managerͷڠྗ͸ෆՄܽʂ (ࣗ෼ͷManager͕આಘͰ͖ͳ͍ͳΒɺ·ͩͦͷ࣌Ͱ͸ͳ͍ͷ͔΋ʁ 2೥લɿॻ੶ “Staff Engineer” ͱͷग़ձ͍ Ikuo ౰࣌ͷManager ͔͔͔͔͘͘͠͡Ͱɺ ΅͘΋Promotion໨ࢦ͍ͨ͠ʂ
34. 34 • Promotion Packet(Self Assessment) • Manager ͱ͍ͬ͠ΐʹετʔϦʔΛߟ͑ɺSelf Assessmentʹ൓ө • ”Staff Project”: ঢ֨ͷ͖͔͚ͬͷϓϩδΣΫτ͸͋ͬͨʁ • Ikuoͷ৔߹: ΦϯϥΠϯ޿ࠂ഑৴γεςϜͷϦΞʔΩςΫνϟઃܭ • ݁Ռɿ̍೥͔͔ͬͨ(νϟϯε͸൒೥ʹ̍౓) • ͭ·Γ̍౓ Reject ͞Ε͍ͯΔɻظ଴͗͢͠ͳ͍ • ͕ͩٸ͗͗ͯ͢΋ྑ͘ͳ͔ͬͨͩΖ͏ • ͱ͖Λ͍͍ͩͨಉͯ͘͡͠ɺ”Staff Engineer” ͷλΠτϧ੍͕ఆ ̍೥લɿTry & Promotionʂ 2-2. ࣮ྫɿIkuoͷ৔߹ ࣗݾΞϐʔϧ͕(ry
35. 35 • Promotion Packet(Self Assessment) • Manager ͱ͍ͬ͠ΐʹετʔϦʔΛߟ͑ɺSelf Assessmentʹ൓ө • ”Staff Project”: ঢ֨ͷ͖͔͚ͬͷϓϩδΣΫτ͸͋ͬͨʁ • Ikuoͷ৔߹: ΦϯϥΠϯ޿ࠂ഑৴γεςϜͷϦΞʔΩςΫνϟઃܭ • ݁Ռɿ̍೥͔͔ͬͨ(νϟϯε͸൒೥ʹ̍౓) • ͭ·Γ̍౓ Reject ͞Ε͍ͯΔɻظ଴͗͢͠ͳ͍ • ͕ͩٸ͗͗ͯ͢΋ྑ͘ͳ͔ͬͨͩΖ͏ • ͱ͖Λ͍͍ͩͨಉͯ͘͡͠ɺ”Staff Engineer” ͷλΠτϧ੍͕ఆ ̍೥લɿTry & Promotionʂ 2-2. ࣮ྫɿIkuoͷ৔߹ ϓϩϞʔγϣϯ΋ΩϟϦΞΞοϓ΋ खஈͰ͋ͬͯ໨తͰ͸ͳ͍ ͜Ε͔ΒԿΛ੒͔͢ʁ͕͍ͩ͡ ࣗݾΞϐʔϧ͕(ry
36. ͲΜͳ࢓ࣄ 03 …Λ͍ͯ͠Δʁ
37. 37 3-1. ࣌ؒͷ࢖͍ํ -- िؒεέδϡʔϧ Ikuoͷ͹͍͋ Ikuo Architect Tech Lead staff archetypes ࢓ࣄ֎ MTG ूத 1 on 1 ໘઀
38. 38 3-1. ࣌ؒͷ࢖͍ํ -- िؒεέδϡʔϧ Ikuoͷ͹͍͋ɿIkuoྲྀ ࣌ݟࣜ… Category Hours % A. ݸਓ࡞ۀ(࣮૷/෼ੳ) 5.0 22.2% B. ઃܭ/ϨϏϡʔ(ઃܭ/υΩϡϝϯτ/ϨϏϡʔ) 9.0 40.0% C. ਪਐɾӡ༻(Working Group/ӡ༻) 4.5 20.0% D. ૊৫ɾར֐ௐ੔(1on1/Ϧʔυ) 4.0 17.8% Category / Archetype A. B. C. D. ྨࣅ౓% Tech Lead 20% 30% 35% 15% 85.0% Architect 10% 50% 20% 20% 87.8% Solver 55% 25% 10% 10% 67.2% Right Hand 5% 10% 35% 50% 52.8% Ikuo Summarized by ChatGPT 5-thinking Similarity by ChatGPT 5-thinking
39. 39 1-3: Ikuoྲྀ ΞʔΩλΠϓ࣌ݟࣜ A.ݸਓ࡞ۀ(࣮૷/෼ੳ) ~22% • σʔλ෼ੳ • ΍ΕΔΤϯδχΞ͕গͳ͍ ˍ IkuoͷಘҙྖҬ • ઃܭ΍ϨϏϡʔ࣌ɺσʔλΛ൑அࡐྉͱͯ͠ఏڙ • ίʔσΟϯά • ׂ߹͸গͳ͍͕ɺखΛ཭͞ͳ͍Α͏ʹ͍ͯ͠Δ B.ઃܭ/ϨϏϡʔ ~40% • ίʔυϨϏϡʔɿґཔ͸ͳΔ΂͘࠷༏ઌ • ૊৫શମͷεϐʔυ/඼࣭ͷఈ্͛ • (ཪ໨త) ৴པஷۚͷ֫ಘ • ํ਑ࡦఆͷυΩϡϝϯτɺઃܭ • ٴͼͦΕΒͷϨϏϡʔ C.ਪਐɾӡ༻ ~20% • Department Working Group • ઌਐٕज़ͷௐࠪ,దԠ • ࠓ͸΋ͪΖΜAI/Agentic Coding • ࢓૊ΈΛ੔͑ΔΑΓ͸৘ใڞ༗͕ϝΠϯ C.૊৫ɾར֐ௐ੔ ~18% • 1on1 • Group಺ओཁϝϯόʔɿϝϯλʔ • ্࢘ɺάϧʔϓ಺ͷϚωδϟʔɿϝϯςΟʔ • ϦʔμʔγοϓMTG • ച্ਐḿ΍ઓུڞ༗ 3-2. ओͳ࢓ࣄ
40. 40 ࢓ࣄָ͍͠ʁͭΒ͍ʁ ͨͷ͠(΢ο… ͭΒ… …େมʂʂָ͍͠οʂ ࢓ࣄ͸͍ͨ΁Μɺ͕͍͔ۙͳ… 3-3. ָ͠͞ɺ΍Γ͕͍ɺେม͞
41. 41 ̍/3 3-3. ָ͠͞ɺ΍Γ͕͍ɺେม͞ νʔϜΛ཭ΕΔ
42. 42 ̍/3 νʔϜΛ཭ΕΔ • ΑΓ޿͍ྖҬ ΁ͷίϛοτ • ਂ͍ूத࣌ؒ Λ֬อ 🧭 Why? 🛠 How? 🎯 So What? • νʔϜఆྫ(εΫ ϥϜΠϕϯτͳ Ͳ)͔Βଔۀ • νʔϜλεΫ௚ ΞαΠϯΛࢭΊ Δ • ࣗ෼ͷ࢓ࣄΛࣗ෼Ͱ ܾΊΔ • ࡋྔˢɺϓϨο γϟʔ↑ • ௚઀తʹνʔϜͷ࢓ ࣄʹख/ޱΛग़ͤΔ λΠϛϯά͕ݮͬͨ IC ͱͯ͠ಇ͖࢝Ί͔ͯΒॳɻେ͖ͳมԽͰָ͍͠ʂͦͯ͠େม… 3-3. ָ͠͞ɺ΍Γ͕͍ɺେม͞
43. 43 ̎/3 େྔͷ৘ใΛॲཧ͢Δ 3-3. ָ͠͞ɺ΍Γ͕͍ɺେม͞
44. 44 • Opportunity/ SupportͰ͖Δͱ ͜ΖΛ୳͢ • ඞཁͳৄࡉΛ཈͑ ͭͭɺେہతͳ໨ ઢΛ࣋ͭ 🧭 Why? 🛠 How? 🎯 So What? • ฒྻ౓Λ্͛ɺؔΘΔϓϩδΣΫ τͷ਺Λ૿΍͢ • ෯ͱਂ͞ͷ୳ࡧ༏ઌॱҐ • ઓུϨϕϧͷ၆ᛌ͔Βɺ࣮૷Ϩ ϕϧͷৄࡉ΁ • Ͳͷఔ౓ͷച্/γΣΞΛ͍ͭ ·Ͱʹʁ • →Ͳͷఔ౓࣌ؒΛ͔͚Δͷ͔ʁ ͍ͭ΍Δͷ͔ʁͲ͏΍Δͷ͔ʁ • ݱ৔Ͱٞ࿦ɾϨϏϡʔͰ͖ Δ࣮૷ΧϯΛอ࣋ͨ͠· ·ɺDivHeadϨϕϧͱձ࿩Ͱ ͖Δେہ؍ • ίϯςΩετεΠονͷ૿େ • ܇࿅ • AI࣌୅ͷ࢓ࣄͷ࢓ํͱ ΋Ϛον͍ͯ͠Δ? ઓུϨϕϧͷෆ࣮֬ੑͱ࣮૷ϨϕϧͱΛͭͳ͙ɻେม… ̎/3 େྔͷ৘ใΛॲཧ͢Δ 3-3. ָ͠͞ɺ΍Γ͕͍ɺେม͞
45. 45 ̏/3 ৴པ͕͢΂ͯʂ 3-3. ָ͠͞ɺ΍Γ͕͍ɺେม͞
46. 46 • “ҕ೚͞ΕͨݖݶͰӨڹ ྗΛൃش͢Δɻ” • ”AuthorityΛआΓΔ” ͸ ৗʹ͸ൃಈͰ͖ͳ͍ • ӨڹྗΛൃش͢Δͱ͖ ͸ɺجຊࣗ෼ͷ৴པஷ ۚΛ࢖͏ 🧭 Why? 🛠 How? 🎯 So What? • ·ͣ૬खΛ৴པͤΑ • ࿩͸ͦΕ͔Βͩ • ઌʹGive • αϙʔτ/Πωʔϒ ϦϯάͰ৴པஷۚ Λ૿΍͢ • צҧ͍͔Βͷ୤٫ • ʮ୭΋๻ͷݴ͏͜ ͱʹฉࣖ͘Λ࣋ͨ ͳ͍…ʯ • ৴པ͕͋ͬͯͦ͜ɺ ΑΓେ͖ͳൣғΛಈ ͔ͤΔ ̏/3 ৴པ͕͢΂ͯʂ ͜ͷ͋ͱ঺հ͢Δࣦഊ͸΄ͱΜͲ͜Ε… 3-3. ָ͠͞ɺ΍Γ͕͍ɺେม͞
47. 47 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍͔͘ͳ͔ͬͨ͜ͱ — ̍/3 Ikuo νʔϜؒͰ࿩ܾͯ͠ΊΒΕͳ͍ͳΒɺ ΅͕͘ Group TL ͱܾͯ͠ΊΔΑʂ ന೤͢Δٞ࿦… XXX͕͍͍ઃܭͩ ͍΍ɺYYYͩ ….. …. 👊୭͕͓લΛGroup TLʹͨ͠ͷʁ 👊ͦΕɺνʔϜTL͕͍Δҙຯͳ͘ͳ͍ʁ
48. 48 Ikuo νʔϜؒͰ࿩ܾͯ͠ΊΒΕͳ͍ͳΒɺ ΅͕͘ Group TL ͱܾͯ͠ΊΔΑʂ XXX͕͍͍ઃܭͩ ͍΍ɺYYYͩ ….. …. 👊୭͕͓લΛGroup TLʹͨ͠ͷʁ 👊ͦΕɺνʔϜTL͕͍Δҙຯͳ͘ͳ͍ʁ ͝΋ͬͱ΋ʂʂ 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍͔͘ͳ͔ͬͨ͜ͱ — ̍/3
49. 49 Ikuo νʔϜؒͰ࿩ܾͯ͠ΊΒΕͳ͍ͳΒɺ ΅͕͘ Group TL ͱܾͯ͠ΊΔΑʂ XXX͕͍͍ઃܭͩ ͍΍ɺYYYͩ ….. …. 👊୭͕͓લΛGroup TLʹͨ͠ͷʁ 👊ͦΕɺνʔϜTL͕͍Δҙຯͳ͘ͳ͍ʁ ࣦഊ — ࣗ෼͕ܾΊΔ(͜ͱʹݻࣥ͢Δ) ԿނʁɿܾΊΔ͜ͱ͕࢓ࣄͱצҧ͍͍ͯͨ͠… ରԠɿఫճ → αϙʔτ͕ࣗ෼ͷத৺ۀ຿Ͱ͋Δ͜ͱΛ໌ࣔ 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍͔͘ͳ͔ͬͨ͜ͱ — ̍/3 ”๻͕ܾΊΔΑʂ”
50. 50 Ikuo νʔϜؒͰ࿩ܾͯ͠ΊΒΕͳ͍ͳΒɺ ΅͕͘ Group TL ͱܾͯ͠ΊΔΑʂ XXX͕͍͍ઃܭͩ ͍΍ɺYYYͩ ….. …. 👊୭͕͓લΛGroup TLʹͨ͠ͷʁ 👊ͦΕɺνʔϜTL͕͍Δҙຯͳ͘ͳ͍ʁ (ͳΓͨͯͷࠒͷ࿩͕ͩ) ΠΩ͍ͬͯͨɺͱ͔͠ݴ͍Α͏͕ͳ͍… 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍͔͘ͳ͔ͬͨ͜ͱ — ̍/3 ”๻͕ܾΊΔΑʂ”
51. 51 Ikuo ͪΐͬͱ·ͬͯɻ ͜͏͜͏͜͏͍͏ཧ༝ͰϠό͍Ͱ͢ɻ ࢲͷݖݶͰϦδΣΫτ͠·͢ A͕Αͦ͞͏ͩ ͦ͏ͩͶɺAͩ ….. …. Α͠ɺ͜ͷҊͰ͍͜͏ʂ 👊ԿͷݖݶͩΑ… 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍͔͘ͳ͔ͬͨ͜ͱ — ̎/3
52. 52 Ikuo ͪΐͬͱ·ͬͯɻ ͜͏͜͏͜͏͍͏ཧ༝ͰϠό͍Ͱ͢ɻ ࢲͷݖݶͰϦδΣΫτ͠·͢ A͕Αͦ͞͏ͩ ͦ͏ͩͶɺAͩ ….. …. 👊ԿͷݖݶͩΑ… Α͠ɺ͜ͷҊͰ͍͜͏ʂ ͝΋ͬͱ΋śőőʂʂ 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍͔͘ͳ͔ͬͨ͜ͱ — ̎/3
53. 53 Ikuo ͪΐͬͱ·ͬͯɻ ͜͏͜͏͜͏͍͏ཧ༝ͰϠό͍Ͱ͢ɻ ࢲͷݖݶͰϦδΣΫτ͠·͢ A͕Αͦ͞͏ͩ ͦ͏ͩͶɺAͩ ….. …. 👊ԿͷݖݶͩΑ… Α͠ɺ͜ͷҊͰ͍͜͏ʂ ࣦഊ — ϦδΣΫτ Կނʁ ɿݖݶͳͲແ͍… ରԠɿࣦഊΛڐ༰ ”͜Ε͸Ϡό͍” ͷᮢ஋ΛԼ͛Δɻ → ҙࢥܾఆͷڧ౓(ඇՄٯ౓ͱӨڹظؒ)Λ֬ೝ͠ɺܾఆΛαϙʔτ͢Δ 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍͔͘ͳ͔ͬͨ͜ͱ — ̎/3 ΞΠσΟΞͷδΣΫτ
54. 54 Ikuo ݟͯݟͯʂ ϐοΧϐΧͷΨΠυϥΠϯΛ࡞ͬͨΑʂʂ ࠷ߴ͔ͩΒ͜ΕͰಇ͜͏ʂ ~~ ίϝϯτ ~~ ίϝϯτ ~~ ίϝϯτ … … … 📁 📁 📁 📁 📁 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍͔͘ͳ͔ͬͨ͜ͱ — ̏/3
55. 55 Ikuo ݟͯݟͯʂ ϐοΧϐΧͷΨΠυϥΠϯΛ࡞ͬͨΑʂʂ ࠷ߴ͔ͩΒ͜ΕͰಇ͜͏ʂ ~~ ίϝϯτ ~~ ίϝϯτ ~~ ίϝϯτ … … … 📁 📁 📁 📁 📁 ແʂʂ 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍͔͘ͳ͔ͬͨ͜ͱ — ̏/3
56. 56 Ikuo ݟͯݟͯʂ ϐοΧϐΧͷΨΠυϥΠϯΛ࡞ͬͨΑʂʂ ࠷ߴ͔ͩΒ͜ΕͰಇ͜͏ʂ ~~ ίϝϯτ ~~ ίϝϯτ ~~ ίϝϯτ … … … 📁 📁 📁 📁 📁 ࣦഊ — ಠΓΑ͕Γ Կނʁɿ”पғΛר͖ࠐΜͰ” Өڹ ରԠɿ෧ҹ → ੍ఆ͔࣌ΒओཁϝϯόʔΛר͖ࠐΉ ͍͔ʹਓΛר͖ࠐΉ͔ɺ͕࢓ࣄͷΩϞ 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍͔͘ͳ͔ͬͨ͜ͱ — ̏/3 ಠΓΑ͕ΓΨΠυϥΠϯ
57. 57 Ikuo ͪΐͬͱ·ͬͯɻ ͜͏͜͏͜͏͍͏ཧ༝ͰϠό͍Ͱ͢ɻ ࢲͷݖݶͰ(ry A͕Αͦ͞͏ͩ ͦ͏ͩͶɺAͩ ….. …. Α͠ɺ͜ͷҊͰ͍͜͏ʂ σʔλ͕ 10 ഒʹͳͬͯ΋଱͑ΒΕΔ ΞʔΩςΫνϟΛ࡞ΔͧʂPJ 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍ͬͨ͘͜ͱ — ̍/2
58. 58 Ikuo ͪΐͬͱ·ͬͯɻ ͜͏͜͏͜͏͍͏ཧ༝ͰϠό͍Ͱ͢ɻ ࢲͷݖݶͰ(ry A͕Αͦ͞͏ͩ ͦ͏ͩͶɺAͩ ….. …. Α͠ɺ͜ͷҊͰ͍͜͏ʂ σʔλ͕ 10 ഒʹͳͬͯ΋଱͑ΒΕΔ ΞʔΩςΫνϟΛ࡞ΔͧʂPJ Ͱ͸ͳ͍ʂʂ 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍ͬͨ͘͜ͱ — ̍/2
59. 59 Ikuo ͳΔ΄Ͳ௅ઓతͩͶʂ ͱ͜ΖͰɺ͍ͭ·Ͱʹ10ഒʹ଱͑ΒΕΔΑ͏ʹ ͳΕ͹͍͍ΜͩΖ͏ʁ ͲΕ͘Β͍͕࣌ؒ࢖͑ͦ͏ʁ A͕Αͦ͞͏ͩ ͦ͏ͩͶɺAͩ ….. …. Α͠ɺ͜ͷҊͰ͍͜͏ʂ σʔλ͕ 10 ഒʹͳͬͯ΋଱͑ΒΕΔ ΞʔΩςΫνϟΛ࡞ΔͧʂPJ 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍ͬͨ͘͜ͱ — ̍/2
60. 60 Ikuo ͳΔ΄Ͳ௅ઓతͩͶʂ ͱ͜ΖͰɺ͍ͭ·Ͱʹ10ഒʹ଱͑ΒΕΔΑ͏ʹ ͳΕ͹͍͍ΜͩΖ͏ʁ ͲΕ͘Β͍͕࣌ؒ࢖͑ͦ͏ʁ A͕Αͦ͞͏ͩ ͦ͏ͩͶɺAͩ ….. …. Α͠ɺ͜ͷҊͰ͍͜͏ʂ ޿ࠂ͕ 10 ഒʹͳͬͯ΋଱͑ΒΕΔ ΞʔΩςΫνϟΛ࡞ΔͧʂPJ 👍 ໰͍ΛཱͯΔ ϝϯόʔ͸ࢦࣔ͞Εͳͯ͘ϋοϐʔ Ikuo͸৴པஷۚΛͨΊͭͭ੒௕͕ݟΕͯϋοϐʔ 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍ͬͨ͘͜ͱ — ̍/2 ࢦࣔΑΓ໰͍
61. 61 Infra Alert, Incident Inquiry Team ػೳ࣮૷ɺΠϯϑϥɺίετ࡟ݮ ͞·͟·ͳґཔ͕෣͍ࠐΉνʔϜ… …PBL͕௕େԽɺ Lead Time ͕௕͘ͳΓ͕ͪ 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍ͬͨ͘͜ͱ —̎/2
62. 62 Infra Alert, Incident Inquiry Team ΩϡʔʹೖΔલʹ ଧͪฦ͢ 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍ͬͨ͘͜ͱ —̎/2 Ikuo
63. 63 Infra Inquiry 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍ͬͨ͘͜ͱ —̎/2 Alert, Incident Team ΩϡʔʹೖΔલʹ ଧͪฦ͢ Ikuo 👍 Πϯλʔηϓτ ґཔओ͸ૣ͘࢓ࣄ͕ऴΘͬͯϋοϐʔ νʔϜ͸Πϯλϥϓτ͕ݮͬͯϋοϐʔ(Invisible͕ͩ… Ikuo ͸खΛಈ͔͢ޱ࣮͕Ͱ͖ͯϋοϐʔ
64. 64 Infra Inquiry 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ͏·͍ͬͨ͘͜ͱ —̎/2 Alert, Incident Team ΩϡʔʹೖΔલʹ ଧͪฦ͢ Ikuo ͨͩ͠ɿ όϥϯεΛऔΒͳ͍ͱແݶࡶ༻ฤʹ…
65. 65 3-4. ࣄྫɿ͏·͍ͬͨ͘͜ͱɺ͍͔ͳ͔ͬͨ͜ͱ ڞ௨͢Δ՝୊ͱֶͼ ՝୊ɿ ➡ܾఆݖ͕͋Δ͔Α͏ͳৼΔ෣͍͸൓ൃΛੜΉ͚ͩ ✓ͦ΋ͦ΋ Staff+ ໾৬ʹ Authority ͸ͳ͍ ✓৴པஷۚͰ࢓ࣄΛ͍ͯ͠Δ͜ͱΛ๨Εͯ͸ͳΒͳ͍ Lesson Learnt: 1. Title ΍आΓ෺ͷ Authority ΛཔΔͳɻ৴པΛಘΑɻ 2. ϦδΣΫτ͢ΔͳɻαϙʔτͤΑɻ 3. ࣗ෼͕ܾΊΔඞཁ͸ͳ͍ɻΑΓྑ͍ܾఆ΁ಋ͚ɻ “Authority͕ͳ͍”͜ͱΛड͚ೖΕ͔ͯΒɺ ৺ཧతෛՙ΋͍ͩͿݮͬͨ
66. 04 ͜Ε͔Β …ICͱͯ͠ੜ͖࢒Δʹ͸ʁ
67. 67 4-1. AIʹΑΔมԽͱ Job Security Q. ͜Ε͔Β Staff + ͸Ͳ͏ͳΔͱࢥ͍·͔͢ʁ Ұ൪੣࣮ͳଶ౓… Ͱ͕͢ɺࠓ೔͸ٞ࿦Λ͠ʹདྷ·ͨ͠ A. Θ͔Βͳ͍ ~ ऴ ~ ͱMartin Fowler ͕ݴ͍ͬͯΔ (஫ɿLLMͱιϑτ΢ΣΞ։ൃʹؔ͢Δߟ࡯)
68. 68 4-1. AIʹΑΔมԽͱ Job Security ௕ظ (10೥~) ୹ظ (਺೔) (਺ϲ݄~൒೥) (Ұ೥~਺೥) ϛΫϩ ϚΫϩ ੈքܦࡁ Staff+ IC IT Software Engineer ೔ຊܦࡁ ࠓ೔ͷٞ࿦ͷ࿮૊Έ ࠓͷࣗ෼ͷ ϙδγϣϯ ௕ظ 㲗 ୹ظ ϚΫϩ 㲗 ϛΫϩ ͷ̐৅ݶͰߟ͑Δ
69. 69 4-1. AIʹΑΔมԽͱ Job Security ௕ظ (10೥~) ୹ظ (਺೔) (਺ϲ݄~൒೥) (Ұ೥~਺೥) ϛΫϩ ϚΫϩ ੈքܦࡁ ࠓͷࣗ෼ͷ ϙδγϣϯ Staff+ IC IT Software Engineer ೔ຊܦࡁ ྺ࢙తࣄ࣮΍ ܦࡁֶ Input Input ࠓ೔ͷٞ࿦ͷ࿮૊Έ IC, Staff + ͷ ਺೥εύϯͷมԽ ͜ͷ΁ΜΛਪଌɾٞ࿦͍ͨ͠ ܦݧͱมԽͷ؍ଌ ٴͼͦͷඍ෼
70. 70 4-1. AIʹΑΔมԽͱ Job Security ௕ظxϚΫϩ — λεΫཧ࿦ Acemoglu, D., & Restrepo, P. (2019). "Automation and New Tasks" • ࢓ࣄ = “λεΫͷଋ” • ࢿຊ(ػց,AI) ͕࣮ߦՄೳͳλεΫ — Ұൠతʹఆܕ(Routine) • ࿑ಇऀ͕ൺֱత༏ҐΛ࣋ͭλεΫ — Ұൠతʹ൱ఆܗ(Non-Routine) • “ٕज़ਐา” ͷ࿑ಇधཁ΁ͷӨڹ — ࢿຊͱ࿑ಇऀͷλεΫ෼഑ͷมԽ • ࣗಈԽ(Automation) ᶃ ੜ࢈ੑޮՌʢProductivity Effectʣ • λεΫͷࣗಈԽʹΑΔੜ࢈ੑ޲্ɻ࿑ಇधཁͷؔઅత૿Ճ • Ձ֨௿Լˠधཁ૿Ճˠੜ࢈֦େˠิ׬తλεΫͷधཁ૿Ճ ᶄ ஔ׵ޮՌʢDisplacement Effectʣ • λεΫͷࣗಈԽʹΑΔ࿑ಇधཁͷ௚઀తݮগ • ৽λεΫͷ૑ग़(Creation of New Tasks) ᶅ ෮ؼޮՌʢReinstatement Effectʣ • ࿑ಇऀ͕ൺֱత༏ҐΛ࣋ͭ৽͍͠λεΫͷ૑ग़ɻ࿑ಇधཁͷ௚઀త૿Ճ
71. 4-1. AIʹΑΔมԽͱ Job Security ௕ظxϚΫϩ — 1947-87 ίϯϐϡʔλීٴ ஔ׵ ෮ؼ ෮ؼ ʔ ஔ׵ ɹ ௞ۚ૯ֹ ੜ࢈ੑ ෮ؼͱஔ׵͕όϥϯε ੜ࢈ੑͷ޲্ʹ൐͍ ௞ۚ૯ֹ޲্ ෮ؼ(λεΫ૑ग़)ͷྫ ίϯϐϡʔλͷීٴʹ൐͍ɺ ϓϩάϥϚͱ͍͏৽ͨͳ࢓ࣄ͕૑ग़ Y࣠:ݪ఺͔Βͷྦྷੵ஋ e.g. 47-87೥Ͱ20%ஔ׵(࿑ಇݮগ
72. 4-1. AIʹΑΔมԽͱ Job Security ஔ׵ ෮ؼ ෮ؼ ʔ ஔ׵ ɹ ௞ۚ૯ֹ ੜ࢈ੑ Ұ؏ͯ͠ஔ׵͕༏Ґ ௕ظxϚΫϩ — ྫɿ1987 - 2017 ~ ϩϘοτ, ιϑτ΢ΣΞԽ ෮ؼ - ஔ׵ͷဃ཭͕େ͖͍ͱ ੜ࢈ੑ্͕͕ͬͯ΋ ௞ۚ૯ֹ͕ఀ଺
73. 4-1. AIʹΑΔมԽͱ Job Security ஔ׵ ෮ؼ ෮ؼ - ஔ׵ ɹ ௞ۚ૯ֹ ੜ࢈ੑ Ұ؏ͯ͠ஔ׵͕༏Ґ ੜ࢈ੑ্͕͕ͬͯ΋ ௞ۚ૯ֹ͕ఀ଺ ࣗಈԽͷՃ଎ʹΑΓஔ׵-෮ؼͷόϥϯε่͕ΕΔͱ ࿑ಇऀͷऔΓ෼͕খ͘͞ͳΔ ௿௞ۚԽɺ͋Δ͍͸ޏ༻ࣗମͷ૕ࣦ ௕ظxϚΫϩ — ྫɿ1987 - 2017 ~ ϩϘοτ, ιϑτ΢ΣΞԽ
74. 4-1. AIʹΑΔมԽͱ Job Security ௕ظxϚΫϩ — λεΫཧ࿦ʹΑΔAIͷӨڹ༧ଌ 1.GPTs are GPTsʢEloundou+ 2023ʣ • ถࠃۀछͷ໿80%͕ۀ຿ͷ10%Ҏ্ͰӨڹ(࿐ग़/Exposure … 50%ͷ࣌ؒͰ׬) • LLM୯ମ(α)ͰશλεΫͷ≈15%͕“ಉ඼࣭Ͱߴ଎Խ”ɺLLM+(γ)ͳΒ≈47–56% • ߴ௞ۚ৬΄Ͳ࿐ग़͕ߴ͍܏޲ 2.The Simple Macroeconomics of AIʢAcemoglu 2024ʣ • TFP(Total Factor Productivity) +0.71% ্ݶ10೥/ݶఆతͳӨڹ • ࿐ग़཰͸ߴ͍͕ɺ࿐ग़λεΫͷGDPൺ͕௿͍(4.6%)ͨΊ • ฏۉ௞ۚӨڹ +1.01%ʢ10೥ʣɺࢿຊ෼഑཰͸֦େ • αϒάϧʔϓ͝ͱͰ͸ϚΠφεɺ࠷େ -0.7% • ࿐ग़౓͸ߴ͍͕ɺGDPൺ͕খ͍͞(ྔ͕গͳ͍)ͨΊܰඍͳ༧ଌ 80%ͷۀछ͕10%Ҏ্࿐ग़ 0.8 0.1
75. 75 4-1. AIʹΑΔมԽͱ Job Security ௕ظxϚΫϩ — Key Takeaway 1. λεΫཧ࿦͕ڭ͑Δٕज़มԽͷϝΧχζϜ • ٕज़ਐาͷೋ໘ੑɿλεΫͷࣗಈԽʢஔ׵ʣͱ৽λεΫͷ૑ग़ʢ෮ؼʣ • ॏཁͳͷ͸ஔ׵ͱ෮ؼͷόϥϯε 2. ྺ࢙͕ࣔ͢ܯࠂͱر๬ • 1947-87೥ɿஔ׵ͱ෮ؼͷόϥϯε → ੜ࢈ੑͱ௞ۚͷಉ࣌੒௕ • 1987-2017೥ɿஔ׵Ճ଎ɾ෮ؼݮ଎ → ௞ۚ੒௕ͷఀ଺ • ڭ܇ɿόϥϯε͸ࣗಈతʹ͸อূ͞Εͳ͍ɺҙࣝతͳ౒ྗ(e.g ੓࣏తհೖ)͕ඞཁ 3. AI࣌୅΁ͷ࣮ફతؚҙ — λεΫཧ࿦ʹΑΔ༧ଌ • ৬ۀʹΑͬͯҟͳΔAI࿐ग़౓ʢExposure — AIஔ׵ՄೳͳλεΫͷ౓߹͍ʣ • → AIʹΑΔӨڹ౓߹͍ͷෆۉҰੑ • ϚΫϩͰ͸࿐ग़λεΫͷGDPγΣΞͷ௿͔͞ΒɺӨڹ͸ݶఆతͱ͍͏༧ଌ
76. 76 4-1. AIʹΑΔมԽͱ Job Security ୹ظxϛΫϩ — “ඍ෼” ᶃ ओ؍ʹΑΔมԽ؍ଌ Apr May June July Aug 2025 Mar IntelliJ + Copilot ChatGPT (ίϐϖ) खಈ࣌୅ Ikuo : AI ~ 80 : 20 Cursor AI Agent ࣌୅ Ikuo : AI ~ 50 : 50 Claude Code Agentic Coding ࣌୅ Ikuo : AI ~ 10 : 90 ΋͸΍खͰ ίʔυΛॻ͘͜ͱ͸ ΄ͱΜͲͳ͍ ݸਓతͳ ඇ࿈ଓ͔ͭ ഁյతมԽ
77. 77 4-1. AIʹΑΔมԽͱ Job Security ୹ظxϛΫϩ — “ඍ෼” ᶄ٬؍తࢦඪͷਐḿ • LLM Coding Benchmarks • SWE-bench Verified — Ref • ࣮ࡏGitHubϦϙͷIssueΛ“ࣗ཯Ͱमਖ਼ͯ͠શςετΛ௨͢”ೳྗ • SWE-Bench͔Βղͷ࿐ग़ɺऑ͍ςετɺֶशσʔλԚછΛऔΓআ͍ͨ500݅αϒηοτ • LCB(LiveCodeBench) — Ref • ڝϓϩʢLeetCode / AtCoder / Codeforcesʣͷ৽ن໰୊Λܧଓऩू͠ɺԚછճආΛॏࢹ ͨ͠ίʔυೳྗϕϯν • pass@1 — Ұճͷग़ྗͷΈͰධՁ • GSO(Global Software Optimization) — Ref • LLM Agent ͷੑೳΛධՁ͢Δൺֱత৽͍͠ϕϯν • ࣮ϓϩδΣΫτͷίϛοτཤྺ͔Βநग़ͨ͠102λεΫ/10ίʔυϕʔεͰɺ଎౓࠷దԽ ύονΛࣗಈੜ੒ • ਓؒΤΩεύʔτͷ଎౓޲্ʢ≥95%ʣͱਖ਼͠͞Ͱ൑ఆ(OPT@1)
78. 78 4-1. AIʹΑΔมԽͱ Job Security * Model ൃද೔ʹ֤ϕϯνείΞΛϓϩοτ * ಉ೔ʹෳ਺͋Δ΋ͷ͸TopͷΈ σʔλιʔε ୹ظxϛΫϩ — “ඍ෼” ᶄ٬؍తࢦඪͷਐḿ SWE-bench Verified — Ref LCB(LiveCodeBench) — Ref GSO(Global Software Optimization) — Ref Apr May June July Aug 2025 Mar 2024 Nov
79. 79 4-1. AIʹΑΔมԽͱ Job Security 2024.11 ~ 2025.4 ʹ େ͖ͳਐḿ Agentic Coding͕࣮༻Ϩϕϧ ʹୡͨ͠ͷ΋͓ͦΒ͘͜ͷࠒ σʔλιʔε ୹ظxϛΫϩ — “ඍ෼” ᶄ٬؍తࢦඪͷਐḿ Apr May June July Aug 2025 Mar 2024 Nov * Model ൃද೔ʹ֤ϕϯνείΞΛϓϩοτ * ಉ೔ʹෳ਺͋Δ΋ͷ͸TopͷΈ
80. 80 4-1. AIʹΑΔมԽͱ Job Security * Model ൃද೔ʹ֤ϕϯνείΞΛϓϩοτ * ಉ೔ʹෳ਺͋Δ΋ͷ͸TopͷΈ 2024.11 ~ 2025.4 ʹ େ͖ͳਐḿ Agentic Coding͕࣮༻Ϩϕϧ ʹୡͨ͠ͷ΋͓ͦΒ͘͜ͷࠒ SWE-bench Verified 75% (͓ͦΒ͘·ͩσʔληοτͷ ໰୊΋͋Γͦ͏͕ͩ…) ΑΓෳࡶͳbenchͰ͸ ·ͩ·ͩ৳ͼ͍ͯΔ σʔλιʔε ௚ۙ ͜ΕΒͷbench͸ανΔஹީ (͢Ͱʹे෼ߴ͍) ୹ظxϛΫϩ — “ඍ෼” ᶄ٬؍తࢦඪͷਐḿ Apr May June July Aug 2025 Mar 2024 Nov
81. 81 4-1. AIʹΑΔมԽͱ Job Security ୹ظxϛΫϩ — ॴײ • SWE-Bench Verified 75% ͷҙຯ͢Δͱ͜Ζ • 75% ͷ Github Issue ΛAI୯ಠͰղܾՄೳ • ୺తʹݴͬͯɺ͢Ͱʹେ఍ͷδϡχΞΑΓ΋༏ल ίʔσΟϯά͚ͩͳΒ΅͘ΑΓ… • ෳࡶ͔ͭن໛ͷେ͖͍໰୊͸·ͩۤखɺ͕ͩ… • ओͳϘτϧωοΫ͸ Context Window ͱ૝૾͞ΕΔ͕ɺ֦େ͍ͯ͠Δ • e.g. claude-sonnet-4-20250514[1m] … ~30K/LOC, 100files? microservice·Δ·Δऩ·Δ • ৽ػೳͷ࣮૷Ͱ͋Ζ͏ͱɺଟ͘͸طଘͷιϦϡʔγϣϯͷ૊Έ߹Θͤ • ໰୊Λখ͘͞ఆٛ͠ɺਖ਼֬ʹ΍Γ͍ͨ͜ͱΛݴޠԽ͢Δೳྗ͕ΩϞ Coding ͦͷ΋ͷ͸͢ͰʹAIͰ΄΅(70%Ҏ্?)୅ସՄೳ ※ ͞Εͨɺͱ͸ݴͬͯͳ͍ɻυϝΠϯʹΑͬͯ΋ҧ͏͠ɺAdoption Rate ͸ ~50%ఔ౓ Ref
82. 82 4-1. AIʹΑΔมԽͱ Job Security ୹ظxϛΫϩ — “࣌ݟࣜ” ͰݟΔAI୅ସ A. ݸਓ࡞ۀ(࣮૷/෼ੳ) ~22% • [Routine/Non-routine] σʔλ෼ੳ — ख๏/σʔλॲཧ͸ஔ͖׵͑ΒΕ͕ͨղऍ͸ࠓͷͱ͜ਓؒ • [Routine] ίʔσΟϯά — 90% Ҏ্ஔ͖׵͑ B. ઃܭ/ϨϏϡʔ ~40% • [Routine] ίʔυ/υΩϡϝϯτϨϏϡʔ — ͍·ͷͱ͜Ζஔ͖׵͑50%ҎԼɺิॿతɻ͕ͩਐΈͦ͏ • [Non-routine] ΞʔΩςΫτɾઃܭ — ͍·ͷͱ͜Ζิॿతɻ C. ਪਐɾӡ༻ ~20% • [Non-routine] ਪਐWG — ઓུͱਓؒͷ૬ख͕த৺ɻਓؒ • [Non-routine] ઌਐٕज़ͷௐࠪ,దԠ — AIͷదԠͷAI୅ସ 🤔 ਓؒ D. ૊৫ɾར֐ௐ੔ ~18% • [Non-routine] 1on1 — ࠷΋ஔ͖׵͕͑஗ͦ͏ • [Non-routine] Ϧʔμʔγοϓઓུ — ઓུࡦఆɾܾఆ͸ࠓͷͱ͜ਓؒ 20 ~ 30% ఔ౓ͷλεΫஔ׵཰ ※ ࣌ؒͱੜ࢈͕ൺྫ͢Δ༁Ͱ͸ͳ͍͕…
83. 83 4-1. AIʹΑΔมԽͱ Job Security ୹ظxϛΫϩ — “࣌ݟࣜ” ͰݟΔAI୅ସ A. ݸਓ࡞ۀ(࣮૷/෼ੳ) ~22% • [Routine/Non-routine] σʔλ෼ੳ — ख๏/σʔλॲཧ͸ஔ͖׵͑ΒΕ͕ͨղऍ͸ࠓͷͱ͜ਓؒ • [Routine] ίʔσΟϯά — 90% Ҏ্ஔ͖׵͑ B. ઃܭ/ϨϏϡʔ ~40% • [Routine] ίʔυ/υΩϡϝϯτϨϏϡʔ — ͍·ͷͱ͜Ζஔ͖׵͑50%ҎԼɺิॿతɻ͕ͩਐΈͦ͏ • [Non-routine] ΞʔΩςΫτɾઃܭ — ͍·ͷͱ͜Ζิॿతɻ C. ਪਐɾӡ༻ ~20% • [Non-routine] ਪਐWG — ઓུͱਓؒͷ૬ख͕த৺ɻਓؒ • [Non-routine] ઌਐٕज़ͷௐࠪ,దԠ — AIͷదԠͷAI୅ସ 🤔 ਓؒ D. ૊৫ɾར֐ௐ੔ ~18% • [Non-routine] 1on1 — ࠷΋ஔ͖׵͕͑஗ͦ͏ • [Non-routine] Ϧʔμʔγοϓઓུ — ઓུࡦఆɾܾఆ͸ࠓͷͱ͜ਓؒ ͔ͭɺஔ͖׵͑ΒΕͨ࣌ؒ͸ଞͷ࢓ࣄͰิర͞Ε͍ͯΔ Claude Codeͷग़ྗ଴ͭؒΨϯϓϥ࡞Ζɺͱ͸ͳΒͳ͍…😭 ͳΜͰʁδΣϰΥϯζͷύϥυοΫεʁ 20 ~ 30% ఔ౓ͷλεΫஔ׵཰ ※ ࣌ؒͱੜ࢈͕ൺྫ͢Δ༁Ͱ͸ͳ͍͕…
84. 84 4-1. AIʹΑΔมԽͱ Job Security ୹ظxϛΫϩ — “࣌ݟࣜ” ͰݟΔAI୅ସ A. ݸਓ࡞ۀ(࣮૷/෼ੳ) ~22% • [Routine/Non-routine] σʔλ෼ੳ — ख๏/σʔλॲཧ͸ஔ͖׵͑ΒΕ͕ͨղऍ͸ࠓͷͱ͜ਓؒ • [Routine] ίʔσΟϯά — 90% Ҏ্ஔ͖׵͑ B. ઃܭ/ϨϏϡʔ ~40% • [Routine] ίʔυ/υΩϡϝϯτϨϏϡʔ — ͍·ͷͱ͜Ζஔ͖׵͑50%ҎԼɺิॿతɻ͕ͩਐΈͦ͏ • [Non-routine] ΞʔΩςΫτɾઃܭ — ͍·ͷͱ͜Ζิॿతɻ C. ਪਐɾӡ༻ ~20% • [Non-routine] ਪਐWG — ઓུͱਓؒͷ૬ख͕த৺ɻਓؒ • [Non-routine] ઌਐٕज़ͷௐࠪ,దԠ — AIͷదԠͷAI୅ସ 🤔 ਓؒ D. ૊৫ɾར֐ௐ੔ ~18% • [Non-routine] 1on1 — ࠷΋ஔ͖׵͕͑஗ͦ͏ • [Non-routine] Ϧʔμʔγοϓઓུ — ઓུࡦఆɾܾఆ͸ࠓͷͱ͜ਓؒ ൒೥ఔ౓ͷ࣌ؒ࣠Ͱ Staff+ͷAI୅ସΛ৺഑͢Δඞཁ͸ͳͦ͞͏ 20 ~ 30% ఔ౓ͷλεΫஔ׵཰ ※ ࣌ؒͱੜ࢈͕ൺྫ͢Δ༁Ͱ͸ͳ͍͕…
85. 85 4-1. AIʹΑΔมԽͱ Job Security ͨͩ͠… (ੜ࢈ੑɿ୯Ґ࣌ؒ͋ͨΓͷ”ੜ࢈ྔ=Ξ΢τϓοτ”ͱҰ୴ஔ͍ͯ) ᶃ: “ੜ࢈ੑ” ͷ֨ࠩ͸޿͕Δ͹͔Γɻ Junior or Senior Ͱ͸ͳ͘ɺ AI or Non-AI ͕෼ਫྮ ʹΞοϓσʔτ͠ଓ͚Δඞཁ͋Γ ᶄ: ඍ෼ʹΑΔ༧ଌ͸ඇ࿈ଓͳมԽʹ੬ऑɻ ͭͶʹ࠷৽ͷಈ޲Λ؍ଌ͢Δ
86. 86 4-2. ࠓޙͷ໾ׂมԽ த௕ظ x ϛΫϩɿStaff +΁ͷӨڹ༧ଌ — ̏ͭͷγφϦΦ ൵؍γφϦΦ — ஔ׵ >>> ෮ؼ த༱γφϦΦ — ஔ׵ > ෮ؼ ָ؍γφϦΦ — ஔ׵ ≦ ෮ؼ
87. 87 4-2. ࠓޙͷ໾ׂมԽ ൵؍ — ”ஔ׵ͷ೾”ɿ༧ଌൃੜ཰ 20-25% ϙδγϣϯ60-70%ݮɻAI ͷΦʔέετϨʔγϣϯͱϏδωεదԠ͕ϝΠϯ த༱ — “νʔϜͻͱΓ”ɿ༧ଌൃੜ཰ 45-50% ϙδγϣϯ20-30%ݮɻAIͰੜ࢈ੑ̏~̐ഒɺҰਓͰ̍ͭͷνʔϜฒͷग़ྗ ָ؍ — “৽ͨͳ൶”ɿ༧ଌൃੜ཰ 20-30% ϙδγϣϯ૿ɻϓϩδΣΫτ૿Ͱधཁ૿ɺ৽ͨͳઐ໳ྖҬͰ໾ׂ֦େ த௕ظ x ϛΫϩɿStaff +΁ͷӨڹ༧ଌ — ̏ͭͷγφϦΦ
88. 88 4-2. ࠓޙͷ໾ׂมԽ ൵؍ — ”ஔ׵ͷ೾”ɿ༧ଌൃੜ཰ 20-25% ϙδγϣϯ60-70%ݮɻAI ͷΦʔέετϨʔγϣϯͱϏδωεదԠ͕ϝΠϯ த༱ — “νʔϜͻͱΓ”ɿ༧ଌൃੜ཰ 45-50% ϙδγϣϯ20-30%ݮɻAIͰੜ࢈ੑ̏~̐ഒɺҰਓͰ̍ͭͷνʔϜฒͷग़ྗ ָ؍ — “৽ͨͳ൶”ɿ༧ଌൃੜ཰ 20-30% ϙδγϣϯ૿ɻϓϩδΣΫτ૿Ͱधཁ૿ɺ৽ͨͳઐ໳ྖҬͰ໾ׂ֦େ Staff+ ͷຊ࣭͸ෆ࣮֬ੑͷ؅ཧͱ૑଄తͳ໰୊ղܾ χϯήϯͷྖ෼ͱͯ͠࢒Γ΍͍͢(ͩΖ͏… த௕ظ x ϛΫϩɿStaff +΁ͷӨڹ༧ଌ — ̏ͭͷγφϦΦ
89. 89 4-2. ࠓޙͷ໾ׂมԽ • ൵؍తʢஔ׵ޮՌʼ෮ؼޮՌʣ/ ஔ׵཰50%, ൃੜ֬཰20-25% • AI ͕ίʔυϕʔεશମΛཧղ͠ɺΞʔΩςΫνϟϨϕϧͷҙࢥܾఆ·Ͱࢧԉ • ඞཁਓ਺͸ݱࡏͷ30-40%·Ͱݮগ͠ɺٕज़຋༁ऀతͳੑ࣭͕ڧ·Δ • AI ͷΦʔέετϨʔγϣϯ͓ΑͼϏδωεదԠ͕ओཁλεΫ • த༱ʢ෦෼తόϥϯεʣ/ ஔ׵཰40%, ൃੜ֬཰45-50% • AI͸ڧྗͳิॿπʔϧͱͯ͠ػೳɺਓؒͷ൑அͱ૑଄ੑ͕ґવͱͯ͠த֩తՁ஋ • ੜ࢈ੑ͕2-3ഒ޲্͠ɺҰਓͰैདྷͷ3-4ਓ෼ͷٕज़తӨڹྗΛൃش • ૯਺͸20-30%ݮগ͢Δ͕ɺ࢒ͬͨ໾ׂͷॏཁੑͱใु͸େ෯ʹ૿େ • ָ؍తʢஔ׵ޮՌ ≦ ෮ؼޮՌʣ/ ஔ׵཰30%, ൃੜ֬཰20-30% • AI͕৽ͨͳՄೳੑΛ։͖ɺ͜Ε·Ͱٕज़తʹෆՄೳͩͬͨϓϩδΣΫτ͕࣮ݱՄೳʹ • AIڠௐΞʔΩςΫτͳͲ৽ઐ໳ྖҬ͕ੜ·Εɺ໾ׂ͕֦େ • ҟྖҬͷڮ౉͠ͱෆ࣮֬ੑ؅ཧ͕ɺAI࣌୅ͷத֩తՁ஋ͱཱͯ֬͠ Staff+ ͷຊ࣭͸ෆ࣮֬ੑͷ؅ཧͱ૑଄తͳ໰୊ղܾ χϯήϯͷྖ෼ͱͯ͠࢒Γ΍͍͢(ͩΖ͏… Ծఆɿ • ݱࡏஔ׵཰ 20% • ೥ؒ੒௕཰ • ൵؍: 50% • த༱: 30% • ఍߅܎਺: 0.6 • Ծʹೳྗతʹஔ׵Ͱ͖ͯ΋ɺ 60%͸౉͞ͳ͍બ୒Λ͢Δ By Claude Opus 4.1 த௕ظ x ϛΫϩɿStaff +΁ͷӨڹ༧ଌ — ৄࡉ
90. 90 5. InnovatorʢΠϊϕʔλʔʣ 6. OrchestratorʢΦʔέετϨʔλʔʣ ໾ׂɿ • AIͷՄೳੑΛϢʔβʔՁ஋ʹม׵͠ɺֵ৽తͳ ੡඼ػೳΛઃܭ • ػցֶशͷಛੑΛਂ͘ཧղ͠ɺϏδωεΰʔϧ ͱٕज़ͷՍ͚ڮͱͳΔ ಛ௃ɿ • ϢʔβʔͷજࡏχʔζΛൃݟͯ͠AIͰղܾࡦΛ ૑ग़ • ϓϩτλΠϐϯά͔Βຊ൪ಋೖ·ͰɺAIػೳͷ ඼࣭ͱUXΛ୲อ • ࢢ৔τϨϯυͱAIٕज़ͷਐԽΛৗʹ೺Ѳ͠ɺڝ ૪༏ҐΛߏங ໾ׂɿ • ։ൃ૊৫શମͷAIπʔϧ׆༻Λઃܭɾਪ ਐ͠ɺΤϯδχΞͷੜ࢈ੑΛ࠷େԽ • ਓؒͱAIͷ࠷దͳڠಇϞσϧΛߏங ಛ௃ɿ • Claude౳ͷ։ൃࢧԉAIΛ૊Έ߹Θͤͨޮ ཰తͳϫʔΫϑϩʔΛઃܭ • AIੜ੒ίʔυͷ඼࣭ج४ͱϨϏϡʔϓϩ ηεΛཱ֬ • νʔϜͷAIϦςϥγʔ޲্ΛϦʔυ͠ɺ ܧଓతͳվળαΠΫϧ΁ 4-2. ࠓޙͷ໾ׂมԽ ৽͍͠Staff +ΞʔΩλΠϓͷఏҊ
91. 91 Solver 導 築 解 補 Tech Lead νʔϜΛಋ͖ɺ ํ޲ੑΛࣔ͢ Architect ٕज़Λઃܭ͠ɺ ࢓૊ΈΛܗͮ͘Δ Right Hand ܦӦ૚Λิࠤ͠ɺ શମΛࢧ͑Δ ෳࡶͳ໰୊Λղ͖΄͙͢ 4-2. ࠓޙͷ໾ׂมԽ ৽͍͠ΞʔΩλΠϓͷఏҊ
92. 92 4-2. ࠓޙͷ໾ׂมԽ ৽͍͠ΞʔΩλΠϓͷఏҊ 補 Solver 導 築 解 Tech Lead νʔϜΛಋ͖ɺ ํ޲ੑΛࣔ͢ Architect ٕज़Λઃܭ͠ɺ ࢓૊ΈΛܗͮ͘Δ Right Hand ܦӦ૚Λิࠤ͠ɺ શମΛࢧ͑Δ ෳࡶͳ໰୊Λղ͖΄͙͢ 革 Innovator AI x ProductͰ ֵ৽తมԽΛ୲͏ 調 Orchestrator ։ൃݱ৔Ͱͷ AI׆༻ΛϦʔυ
93. 93 4-2. ࠓޙͷ໾ׂมԽ طଘλΠϓͷมԽ 補 Definer 導 統 定 Tech Lead νʔϜΛಋ͖ɺ ํ޲ੑΛࣔ͢ Governor ݪଇͱΨΠυϥΠϯͰ AIઃܭΛ ”࣏ΊΔ” Right Hand ܦӦ૚Λิࠤ͠ɺ શମΛࢧ͑Δ ෳࡶͳ໰୊ΛAIʹղ͚ΔܗͰ ఆٛ͢Δ 革 Innovator AI x ProductͰ ֵ৽తมԽΛ୲͏ 調 Orchestrator ։ൃݱ৔Ͱͷ AI׆༻ΛϦʔυ ઃܭͷͨΊͷઃܭɺ ϝλઃܭ΁ ղ͘͜ͱΑΓ΋ɺ ໰୊ͷఆٛͦͷ΋ͷʹՁ஋
94. 94 4-2. ࠓޙͷ໾ׂมԽ த௕ظ x ϛΫϩɿIC΁ͷӨڹ༧ଌ — ̏ͭͷγφϦΦ ൵؍γφϦΦ — ஔ׵ >>> ෮ؼ த༱γφϦΦ — ஔ׵ > ෮ؼ ָ؍γφϦΦ — ஔ׵ ≦ ෮ؼ
95. 95 4-2. ࠓޙͷ໾ׂมԽ ൵؍ — ”ஔ׵ͷཛྷ”ɿ༧ଌൃੜ཰ 30-35% ϙδγϣϯ70-80%ݮɻۓٸରԠ,ཁ݅ೖྗ,ग़ྗݕূ͋Δ͍͸৬छస׵ த༱ — “ੜ࢈ϒʔετ”ɿ༧ଌൃੜ཰ 45-50% ϙδγϣϯ20-30%ݮɻ࣮૷࣌ؒͷେ෯ͳ୹ॖɺAI׆༻ͱग़ྗݕূ ָ؍ — “৽ͨͳ൶”ɿ༧ଌൃੜ཰ 20-25% ϙδγϣϯ૿ɻϓϩδΣΫτ૿ɺٕज़తෳࡶੑ૿ʹΑΔधཁ૿ த௕ظ x ϛΫϩɿIC΁ͷӨڹ༧ଌ — ̏ͭͷγφϦΦ
96. 96 4-2. ࠓޙͷ໾ׂมԽ ൵؍ — ”ஔ׵ͷཛྷ”ɿ༧ଌൃੜ཰ 30-35% ϙδγϣϯ70-80%ݮɻۓٸରԠ,ཁ݅ೖྗ,ग़ྗݕূ͋Δ͍͸৬छస׵ த༱ — “ੜ࢈ϒʔετ”ɿ༧ଌൃੜ཰ 45-50% ϙδγϣϯ20-30%ݮɻ࣮૷࣌ؒͷେ෯ͳ୹ॖɺAI׆༻ͱग़ྗݕূ ָ؍ — “৽ͨͳ൶”ɿ༧ଌൃੜ཰ 20-25% ϙδγϣϯ૿ɻϓϩδΣΫτ૿ɺٕज़తෳࡶੑ૿ʹΑΔधཁ૿ ໌֬ͳλεΫͷ઎ΊΔׂ߹͕ߴ͍ͱɺࣗಈԽ΁ͷ࿐ग़͕ߴ͍ AI׆༻΍ϏδωευϝΠϯཧղɺมԽ΁ͷదԠ͕ΧΪ͔ த௕ظ x ϛΫϩɿIC΁ͷӨڹ༧ଌ — ̏ͭͷγφϦΦ
97. 97 4-2. ࠓޙͷ໾ׂมԽ • ൵؍తʢஔ׵ޮՌʼ෮ؼޮՌʣ/ ஔ׵཰90%, ൃੜ֬཰30-35% • νέοτϕʔεͷ։ൃ࡞ۀͷ90%Ҏ্͕AIʹΑͬͯࣗಈԽ • ૯਺͸70-80%࡟ݮ • ۓٸରԠ΍AI΁ͷཁ݅ೖྗ/ग़ྗ֬ೝͷΑ͏ͳλεΫ͔ɺҧ͏ઐ໳ྖҬ΁ͷ৬छస׵ • த༱ʢ෦෼తόϥϯεʣ/ ஔ׵཰70%, ൃੜ֬཰45-50% • AI͸ڧྗͳิॿ໾ͱͯ͠ػೳ͠ɺجຊ࣮૷͕࣌ؒେ෯୹ॖ • ૯਺͸20-30%ݮগ • AIπʔϧ׆༻ೳྗͱग़ྗݕূεΩϧ >>> ७ਮͳίʔσΟϯάྗ • ָ؍తʢஔ׵ޮՌ ≦ ෮ؼޮՌʣ/ ஔ׵཰60%, ൃੜ֬཰20-25% • ϓϩάϥϛϯάͷຽओԽʹΑΓɺແ਺ͷখن໛ϓϩδΣΫτ͕ര஀ • ٕज़ϝϯλʔ΍ίʔυΩϡϨʔλʔͱͯ͠ɺ৽ͨͳઐ໳ੑͱधཁ͕૑ग़ • ૊৫ͷٕज़తෳࡶੑ૿େʹΑΓɺICΤϯδχΞͷधཁ͕૿Ճ ໌֬ͳλεΫͷ઎ΊΔׂ߹͕ߴ͘ɺࣗಈԽ΁ͷ࿐ग़͕ߴ͍ AI׆༻΍ϏδωευϝΠϯཧղɺมԽ΁ͷదԠ͕ΧΪ͔ Ծఆɿ • ݱࡏஔ׵཰40% • ೥ؒAI੒௕཰ • ൵؍: 50% • த༱: 30% • ఍߅܎਺: 1.0(ஔ͖׵͑ΒΕ Δ΋ͷ͸͢΂ͯஔ͖׵͑Δ) By Claude Opus 4.1 த௕ظ x ϛΫϩɿIC΁ͷӨڹ༧ଌ — ৄࡉ
98. ·ͱΊɿKey Takeaways! 1. Staff Engineer ͱ͸ʁ • ٕज़ࢀ๳ɺManager Ҏ֎ͷ Individual Contributor ͱͯ͠ͷΩϟϦΞ • ̐ͭͷ Archetype — ಋ / ங / ղ / ิ 2. Ͳ͏΍ͬͯͳΔʁ • λΠτϧͷ͋Δձࣾʹ… • Promotion Packet Λ४උ͢Δ/ظ଴͗͢͠ͳ͍ 3. ͲΜͳ࢓ࣄΛ͍ͯ͠Δʁ • Ikuo: ࣮૷/෼ੳ ~22%, ઃܭ/ϨϏϡʔ ~40%, ਪਐ/ӡ༻ ~20%, ૊৫/ར֐ௐ੔ ~18% • Team TL ͔ΒӨڹͷ֦େɺෆ࣮֬ੑͷ؅ཧɺ໰୊ఆٛͱղܾɻAuthority Ͱ͸ͳ͘৴པ 4. ͜Ε͔Βʁ • ̏೥͘Β͍ͷεύϯͰ͸ɺStaff +৬͕ফ͑ڈΔ͜ͱ͸ͳ͍ͩΖ͏ɻͨͩ͠มԽ͸͋Δ • ܧଓతͳΩϟονΞοϓ/Ξοϓσʔτ͕େࣄ
99. Thank you for Your Kind Attention!
100. 100 References • ελοϑΤϯδχΞ ― ϚωδϝϯτΛ௒͑ΔϦʔμʔγοϓ • Blog: Staff Engineer: Leadership beyond the management track • ελοϑΤϯδχΞͷಓ ―༏Εٕͨज़ઐ໳৬ʹͳΔͨΊͷΨΠυ • Acemoglu, Daron, and Pascual Restrepo. 2019. "Automation and New Tasks: How Technology Displaces and Reinstates Labor." Journal of Economic Perspectives 33 (2): 3–30. • GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models • Daron Acemoglu, 2024. "The Simple Macroeconomics of AI," NBER Working Papers 32487, National Bureau of Economic Research, Inc. • LLM Coding Performance Bench • SWE-Bench+: Enhanced Coding Benchmark for LLMs • SWE-Bench Verified • LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code • LiveCodeBench Leaderboard • Datasource of 4-1’s Benchmark Fig • Gathered by ChatGPT5 thinking from above leaderboard • Date refers each model’s published date