TestForge | Aidevops | ๐Ÿ“Š Plogger โœ๏ธ Blog ๐Ÿ“š Docs
plogger

AI DevOps Korea

AI ์„œ๋น„์Šค ๊ฐœ๋ฐœ, ์šด์˜, ์„ฑ๋Šฅ๊ฐœ์„ ์„ ํ•˜๋‚˜์˜ ๋ฃจํ”„๋กœ ์—ฐ๊ฒฐํ•ฉ๋‹ˆ๋‹ค

aidevops.kr์—์„œ LLMOps, RAG, AI Agent, ๊ด€์ธก์„ฑ, ํ‰๊ฐ€, ๋น„์šฉ-์„ฑ๋Šฅ ์ตœ์ ํ™”๋ฅผ ์‹ค์ „ ์šด์˜ ๊ด€์ ์œผ๋กœ ์ •๋ฆฌํ•ฉ๋‹ˆ๋‹ค.

Aidevops ๋ฐ”๋กœ๊ฐ€๊ธฐ LLMOps
๐Ÿค– AI / LLMOps / LLMOps ์šด์˜

LLMOps ์šด์˜

AI / LLMOps ์•ˆ์—์„œ ์„œ๋กœ ์—ฐ๊ฒฐํ•ด ์ฝ๊ธฐ ์ข‹์€ ๊ธ€์„ ์ด ๊ทธ๋ฃน์œผ๋กœ ๋ฌถ์—ˆ์Šต๋‹ˆ๋‹ค. ์ž…๋ฌธ, ๋น„๊ต, ์‹ฌํ™” ํƒ์ƒ‰์„ ํ•œ ํ๋ฆ„์œผ๋กœ ์ด์–ด๊ฐ€๊ธฐ ์ข‹์Šต๋‹ˆ๋‹ค.

์ด ๊ทธ๋ฃน์—๋Š” ํ˜„์žฌ 9๊ฐœ์˜ ๊ธ€์ด ์žˆ์Šต๋‹ˆ๋‹ค.

Start Here

์ด ๊ทธ๋ฃน์—์„œ ๋จผ์ € ์ฝ์œผ๋ฉด ์ข‹์€ ๊ธ€

์—์ด์ „ํŠธ ๋ฉ”๋ชจ๋ฆฌ ์œˆ๋„์šฐ ์˜ˆ์‚ฐ ์„ค๊ณ„ ๊ฐ€์ด๋“œ

AI ์—์ด์ „ํŠธ๋Š” ๋งŽ์ด ๊ธฐ์–ตํ• ์ˆ˜๋ก ์ข‹์•„ ๋ณด์ด์ง€๋งŒ, ์‹ค์ œ ์šด์˜์—์„œ๋Š” ๋ฉ”๋ชจ๋ฆฌ ์˜ˆ์‚ฐ๊ณผ ์š”์•ฝ ๊ทœ์น™์ด ํ’ˆ์งˆ์„ ์ขŒ์šฐํ•ฉ๋‹ˆ๋‹ค.

Group Archive

๊ทธ๋ฃน์— ์†ํ•œ ์ „์ฒด ๊ธ€

์—์ด์ „ํŠธ ๋ฉ”๋ชจ๋ฆฌ ์œˆ๋„์šฐ ์˜ˆ์‚ฐ ์„ค๊ณ„ ๊ฐ€์ด๋“œ

AI ์—์ด์ „ํŠธ๋Š” ๋งŽ์ด ๊ธฐ์–ตํ• ์ˆ˜๋ก ์ข‹์•„ ๋ณด์ด์ง€๋งŒ, ์‹ค์ œ ์šด์˜์—์„œ๋Š” ๋ฉ”๋ชจ๋ฆฌ ์˜ˆ์‚ฐ๊ณผ ์š”์•ฝ ๊ทœ์น™์ด ํ’ˆ์งˆ์„ ์ขŒ์šฐํ•ฉ๋‹ˆ๋‹ค.
#ai#llmops#agents#memory

Responses API์™€ Remote MCP ์‹ค์ „ ๋„์ž… ํฌ์ธํŠธ

๋ชจ๋ธ API๊ฐ€ ๋‹จ์ˆœ ํ…์ŠคํŠธ ์‘๋‹ต์„ ๋„˜์–ด ๋„๊ตฌ ํ˜ธ์ถœ ํ”Œ๋žซํผ์œผ๋กœ ๋ฐ”๋€Œ๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. Responses API์™€ Remote MCP๋ฅผ ์ œํ’ˆ ๊ด€์ ์—์„œ ์–ด๋–ป๊ฒŒ ๋ด์•ผ ํ•˜๋Š”์ง€ ์ •๋ฆฌํ•ฉ๋‹ˆ๋‹ค.
#ai#llmops#mcp#responses-api

LLM ์ปจํ…์ŠคํŠธ ์œˆ๋„์šฐ ์˜ˆ์‚ฐ ์„ค๊ณ„

ํ† ํฐ์„ ๋งŽ์ด ๋„ฃ๋Š” ๊ฒƒ์ด ํ•ญ์ƒ ์ข‹์€ ๊ฒƒ์€ ์•„๋‹™๋‹ˆ๋‹ค. ์‹ค์ „ AI ์„œ๋น„์Šค์—์„œ ์ปจํ…์ŠคํŠธ ์œˆ๋„์šฐ๋ฅผ ์–ด๋–ป๊ฒŒ ์˜ˆ์‚ฐํ™”ํ•˜๊ณ  ์šด์˜ํ•ด์•ผ ํ•˜๋Š”์ง€ ์ •๋ฆฌํ•ฉ๋‹ˆ๋‹ค.
#ai#llm#context-window#llmops

LLMOps ํ”Œ๋žซํผ ์•„ํ‚คํ…์ฒ˜: LLM ๊ธฐ๋Šฅ์„ ์šด์˜ ํ™˜๊ฒฝ์—์„œ ๊ตด๋ฆฌ๋Š” ๋ฐฉ๋ฒ•

์š”์ฒญ ๋ผ์šฐํŒ…, ํ”„๋กฌํ”„ํŠธ ๋ฒ„์ „ ๊ด€๋ฆฌ, ์ถ”์ ์„ฑ, ํด๋ฐฑ ์ „๋žต, ํ‰๊ฐ€ ๋ฃจํ”„, ๋น„์šฉ ์ œ์–ด, ์šด์˜ ์ฑ…์ž„ ๋ถ„๋ฆฌ๋ฅผ ํฌํ•จํ•œ LLMOps ์•„ํ‚คํ…์ฒ˜๋ฅผ ์„ค๋ช…ํ•ฉ๋‹ˆ๋‹ค.
#ai#llmops#observability#deployment#monitoring#cost-control

ํ”„๋กฌํ”„ํŠธ ์—”์ง€๋‹ˆ์–ด๋ง ์šด์˜ ๊ฐ€์ด๋“œ: ๋ฒ„์ „ ๊ด€๋ฆฌ, ํ…Œ์ŠคํŠธ, ์‹คํŒจ ๋ณต๊ตฌ

ํ”„๋กฌํ”„ํŠธ ๊ณ„์•ฝ, ๊ตฌ์กฐํ™” ์ถœ๋ ฅ, ๋ฒ„์ „ ๊ด€๋ฆฌ, ํ‰๊ฐ€, ๋กค๋ฐฑ, ํŒ€ ํ˜‘์—… ํ๋ฆ„์„ ํฌํ•จํ•ด ํ”„๋กฌํ”„ํŠธ ์—”์ง€๋‹ˆ์–ด๋ง์„ ์šด์˜ ๊ด€์ ์œผ๋กœ ์„ค๋ช…ํ•ฉ๋‹ˆ๋‹ค.
#ai#prompt-engineering#structured-output#guardrails#evaluation#hallucination

RAG ํ‰๊ฐ€ ํ”Œ๋ ˆ์ด๋ถ: ์‚ฌ์šฉ์ž๊ฐ€ ์‹ ๋ขฐ๋ฅผ ์žƒ๊ธฐ ์ „์— ๊ฒ€์ƒ‰ ํ’ˆ์งˆ์„ ์ธก์ •ํ•˜๋Š” ๋ฒ•

๋ฌธ์„œ ์ปค๋ฒ„๋ฆฌ์ง€, ๋žญํ‚น ํ’ˆ์งˆ, ๋‹ต๋ณ€ ๊ทผ๊ฑฐ์„ฑ, ์‹คํŒจ ๋ถ„์„, ๋ฆด๋ฆฌ์ฆˆ ๊ฒŒ์ดํŠธ๋ฅผ ํฌํ•จํ•œ RAG ํ‰๊ฐ€ ๋ฐฉ๋ฒ•์„ ์‹ค๋ฌด ๊ธฐ์ค€์œผ๋กœ ์ •๋ฆฌํ•ฉ๋‹ˆ๋‹ค.
#ai#rag#retrieval#evaluation#grounding#vector-db