ME News Report, April 18 (UTC+8), recently, Renuka Kumar and Prashanth Ramagopal from the Cisco team built a multi-agent coordination framework based on LangSmith and LangGraph, designed to simulate real-world software team collaboration. The framework falls under the category of "Agent Engineering," aiming to accelerate the entire software process from requirements to deployment through simulation of engineering team collaboration, not just code generation. The system includes work agents responsible for executing development, testing, debugging, and other tasks, as well as leadership agents responsible for coordination, governance, and providing shared resources and long-term memory. Preliminary practices show that in over 20 debugging workflow pilots, root cause analysis time was reduced by 93% compared to the historical baseline, saving over 200 engineering hours in a single month with 512 sessions; execution time for development workflows was reduced by 65%, mainly due to compression of downstream testing phases. (Source: InFoQ)

View Original

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.

8 Likes

Reward
8
14
1
Share

Comment

Add a comment

GateUser-cbb8cdf5

· 5h ago

Root cause localization reduction of 93% is more noteworthy than a 65% reduction in execution, indicating that diagnosis is more difficult to automate than execution.

View OriginalReply0

GlassDome

· 8h ago

512 conversations save 200 work hours, resulting in approximately 23 minutes saved per session; debugging efficiency truly improves significantly.

View OriginalReply0

PleaseReadTheWhitepaperFirst.

· 12h ago

The pilot data looks good, but the sample size of 20 workflows is too small, awaiting validation with thousands of samples.

View OriginalReply0

FoldedPrivateKey

· 13h ago

Leadership proxies manage memory and governance, this architecture is like giving AI a CTO.

View OriginalReply0

GateUser-7df43e29

· 13h ago

93% is a bit outrageous, but using LangGraph as a coordination layer is indeed much more flexible than hard-coded state machines.

View OriginalReply0

QuantitativeButNotPretentious

· 13h ago

Single-month data looks good, but is the ROI calculated based on working hours or labor costs?

View OriginalReply0

CandleLibrarian

· 13h ago

LangSmith做observability，LangGraph做orchestration，思科选栈很务实

Reply0

Post-RainCandlestickReflection

· 13h ago

Waiting for an open-source implementation; want to see how the leadership's proxy prompt is written.

View OriginalReply0

0xLateComer

· 13h ago

65% reduction in execution time, feels like the flywheel of LLM-generated code + automated testing has started turning.

View OriginalReply0

ResilientGoldfish

· 13h ago

Will debugging proxies and testing proxies blame each other? How does governance proxy arbitrate?

View OriginalReply0

Trending Topics
View More
#
StockTradingChallengeUpTo17000U
15.93M Popularity
#
USIranDraftDeal
288.86K Popularity
#
TradeCFDWinGold
3.03M Popularity
#
HYPEMarketCapSurpassesDOGE
12.64M Popularity
#
PlatinumCardCreatorExclusive
155.8K Popularity

Pinned

Sitemap

Cisco team builds multi-agent coordination framework based on LangSmith and LangGraph

Trending Topics

StockTradingChallengeUpTo17000U

USIranDraftDeal

TradeCFDWinGold

HYPEMarketCapSurpassesDOGE

PlatinumCardCreatorExclusive

Pinned