--- schema: 1 type: video title: >- We Mapped 115 Microservices for Our Coding Agents — Kamalakannan Nandagopal, Postman description: >- Coding agents are great inside one repo. Postman has 115 microservices and thousands of endpoints. Kamalakannan Nandagopal, staff engineer at Postman, shares how Postman made coding agents work across a real distributed system. Docs go stale, skills and MCPs are only as good as their context, and agent memory stays stuck with one person. Borrowing from Postman's go-to-market team, they realized APIs are the context layer, and built an API context graph that maps every service, endpoint, implementation and call, all the way down to databases, with every data point grounded in code or production telemetry. He shares evals built from real PRs and design decisions: big wins on API discovery, redesign and impact assessment, an honest failure when the graph went stale, the finding that 75% of PRs touched APIs, and a 21-page architecture review the agent produced on its own. In this talk: • Why coding agents struggle with distributed systems and microservices • Building an API context graph grounded in code and telemetry • Eval results on API discovery, redesign and impact assessment • Why keeping the graph fresh matters as much as building it SPEAKER Kamalakannan Nandagopal, Staff Engineer, Postman LinkedIn: https://www.linkedin.com/in/kamalakannan-nandagopal-3650818a/ LINKS Postman Context Graph API: https://blog.postman.com/introducing-the-context-graph-api-one-map-of-your-api-ecosystem/ Context Graph docs: https://learning.postman.com/docs/use/context-graph/get-started Postman: https://www.postman.com Postman on X: https://x.com/getpostman CHAPTERS 0:00 Intro 0:32 115 microservices 1:02 From autocomplete to autonomous agents 1:42 Real systems are distributed 2:12 Docs go stale 2:36 Skills, MCPs and agent memory 3:21 What the go-to-market team learned 4:01 APIs are the context layer 4:36 Building the API context graph 5:36 Grounded in code and telemetry 5:56 What the graph revealed 6:25 Evals from real PRs 7:35 Reading the results 8:20 Use case: API discovery 9:05 Use case: API redesign 11:54 Use case: impact assessment 12:29 A failure: stale data 13:38 75% of PRs touch APIs 14:28 A 21-page architecture review 15:03 What's next 15:38 Takeaways Recorded at the AI Engineer World's Fair 2026 in San Francisco. Subscribe for more talks from the engineers building with AI. AI Engineer: https://ai.engineer YouTube: https://www.youtube.com/@aiDotEngineer X: https://x.com/aiDotEngineer LinkedIn: https://www.linkedin.com/company/aidotengineer/ #CodingAgents #APIs #AIEngineer canonical: 'https://liquidgold.cc/watch/k2ClBT4aqAg' markdown: 'https://liquidgold.cc/watch/k2ClBT4aqAg.md' llms: 'https://liquidgold.cc/watch/k2ClBT4aqAg/llms.txt' parent: 'https://liquidgold.cc/channel/aidotengineer.md' channel: aidotengineer updated: '2026-10-23T22:19:30.762Z' language: null children: [] videoId: k2ClBT4aqAg durationSeconds: 1046 publishedAt: '2026-10-08T21:30:19.000Z' searchChunks: - ordinal: 0 text: null startSeconds: 47 endSeconds: 67 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=47' heldReason: rights - ordinal: 1 text: null startSeconds: 67 endSeconds: 80 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=67' heldReason: rights - ordinal: 2 text: null startSeconds: 102 endSeconds: 112 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=102' heldReason: rights - ordinal: 3 text: null startSeconds: 147 endSeconds: 157 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=147' heldReason: rights - ordinal: 4 text: null startSeconds: 157 endSeconds: 169 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=157' heldReason: rights - ordinal: 5 text: null startSeconds: 213 endSeconds: 247 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=213' heldReason: rights - ordinal: 6 text: null startSeconds: 258 endSeconds: 281 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=258' heldReason: rights - ordinal: 7 text: null startSeconds: 291 endSeconds: 336 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=291' heldReason: rights - ordinal: 8 text: null startSeconds: 336 endSeconds: 347 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=336' heldReason: rights - ordinal: 9 text: null startSeconds: 370 endSeconds: 391 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=370' heldReason: rights - ordinal: 10 text: null startSeconds: 403 endSeconds: 458 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=403' heldReason: rights - ordinal: 11 text: null startSeconds: 504 endSeconds: 517 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=504' heldReason: rights - ordinal: 12 text: null startSeconds: 561 endSeconds: 724 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=561' heldReason: rights - ordinal: 13 text: null startSeconds: 724 endSeconds: 735 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=724' heldReason: rights - ordinal: 14 text: null startSeconds: 767 endSeconds: 821 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=767' heldReason: rights - ordinal: 15 text: null startSeconds: 821 endSeconds: 832 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=821' heldReason: rights - ordinal: 16 text: null startSeconds: 876 endSeconds: 886 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=876' heldReason: rights - ordinal: 17 text: null startSeconds: 909 endSeconds: 919 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=909' heldReason: rights - ordinal: 18 text: null startSeconds: 942 endSeconds: 954 watch: 'https://liquidgold.cc/watch/k2ClBT4aqAg?t=942' heldReason: rights --- # We Mapped 115 Microservices for Our Coding Agents — Kamalakannan Nandagopal, Postman Coding agents are great inside one repo. Postman has 115 microservices and thousands of endpoints. Kamalakannan Nandagopal, staff engineer at Postman, shares how Postman made coding agents work across a real distributed system. Docs go stale, skills and MCPs are only as good as their context, and agent memory stays stuck with one person. Borrowing from Postman's go-to-market team, they realized APIs are the context layer, and built an API context graph that maps every service, endpoint, implementation and call, all the way down to databases, with every data point grounded in code or production telemetry. He shares evals built from real PRs and design decisions: big wins on API discovery, redesign and impact assessment, an honest failure when the graph went stale, the finding that 75% of PRs touched APIs, and a 21-page architecture review the agent produced on its own. In this talk: • Why coding agents struggle with distributed systems and microservices • Building an API context graph grounded in code and telemetry • Eval results on API discovery, redesign and impact assessment • Why keeping the graph fresh matters as much as building it SPEAKER Kamalakannan Nandagopal, Staff Engineer, Postman LinkedIn: https://www.linkedin.com/in/kamalakannan-nandagopal-3650818a/ LINKS Postman Context Graph API: https://blog.postman.com/introducing-the-context-graph-api-one-map-of-your-api-ecosystem/ Context Graph docs: https://learning.postman.com/docs/use/context-graph/get-started Postman: https://www.postman.com Postman on X: https://x.com/getpostman CHAPTERS 0:00 Intro 0:32 115 microservices 1:02 From autocomplete to autonomous agents 1:42 Real systems are distributed 2:12 Docs go stale 2:36 Skills, MCPs and agent memory 3:21 What the go-to-market team learned 4:01 APIs are the context layer 4:36 Building the API context graph 5:36 Grounded in code and telemetry 5:56 What the graph revealed 6:25 Evals from real PRs 7:35 Reading the results 8:20 Use case: API discovery 9:05 Use case: API redesign 11:54 Use case: impact assessment 12:29 A failure: stale data 13:38 75% of PRs touch APIs 14:28 A 21-page architecture review 15:03 What's next 15:38 Takeaways Recorded at the AI Engineer World's Fair 2026 in San Francisco. Subscribe for more talks from the engineers building with AI. AI Engineer: https://ai.engineer YouTube: https://www.youtube.com/@aiDotEngineer X: https://x.com/aiDotEngineer LinkedIn: https://www.linkedin.com/company/aidotengineer/ #CodingAgents #APIs #AIEngineer Canonical: https://liquidgold.cc/watch/k2ClBT4aqAg · Channel: [AI Engineer](https://liquidgold.cc/channel/aidotengineer.md) > Search chunk text is held until public-source provenance is verified. ## Document chunks > Missing: no stored document chunks are available.