AI Engineer Code Summit · Tue 10 Nov 2026 · Continental 1–3

Own the stack, build the platform

Utkarsh Shrivastava helped roll out one internal AI assistant to more than 10,000 people at Netflix. Underneath the specifics is a shift: choosing a model is no longer the decision. A company owns the stack below it (the runtime, its data, its context) and the real work is building the platform on top.

Utkarsh Shrivastava, Netflix · 27 min · six lessons · Links and further reading
Lesson 1

The model is no longer the decision

They started by trying to pick a tool, and the choice wouldn't hold still. Frontier labs launched enterprise offerings; open-weight models added enterprise support during the year; domain teams had built their own small agents. While they evaluated one option, new ones kept arriving.

So he stopped treating it as the decision. The AI tool you adopt changes so fast that by the time a contract is signed, the vendor has shipped more. The agent harness, the runtime, can be swapped between open-source and frontier options.

From the notes: his frontier labs included xAI, and “open-weight models with enterprise support” was a real option. The reading: the infrastructure has become something you can own. Run supported open-weight models yourself where data can't leave, use frontier models where the capability is worth it, and keep the harness portable across both. (Research: analysis/enterprise-ai-deployment-options.md.)

Evidence: transcript 00:00:43, 00:02:11

Lesson 2

What you own is below the surface

If the model is swappable, the durable assets are the ones underneath: the harness, and above all the company's own skills and data connectors, which is where he says to spend the time.

“The company context is the key here.”00:02:11

The data was already yours; what changed is that it's all reachable. Before, each department had its own application on its own slice of data. Now one assistant reaches across teams, and turning on another source is a toggle.

From the notes: “company context is the key”. The same conclusion George Fraser reached that afternoon from the data side: own your data and the context around it; the model is the replaceable part.

Evidence: transcript 00:02:11, 00:03:12

Lesson 3

Connecting is cheap; deciding isn't

Owning everything and connecting everything creates the new problem.

“Is another way for the agent to choose wrong.”00:04:40

Asked how many people watched a live MMA event Netflix had just streamed, the assistant said 12.4 million from the analytics connector. Weeks later, after a connector to the live systems was added, the same question got nearly 17 million. Both were right: an average and a peak.

“the tool was never wrong.”00:04:40

The pattern is general: one word means different things to different teams, two systems can both truthfully claim the answer, and connecting everything never settled who owns it. Users who didn't get the answer they expected stopped trusting it fast. That gap is what the platform is for.

Evidence: transcript 00:04:40, 00:06:49

Lesson 4

The platform decides who answers

The platform's first job is deciding which answer is right, and for whom. When sources conflict, don't silently prefer the newest or the oldest: show both, labeled.

“the right thing to do is to label both”00:13:04

Use who is asking: an engineer moving live traffic and a business user planning content want different numbers for the same words. Set the agent's behavior: an interactive agent should ask what you mean; a long-running one should pick and state its assumption. In the Q&A he describes Netflix settling its own case by routing: live systems first, handing off to analytics when that's the better source.

Evidence: transcript 00:13:04, 00:21:38

Lesson 5

The platform makes it hold, across any model

The rest of the platform is guarantees. Access by role first. Then a regression suite that domain teams run on every new connector, skill or plugin, starting from twenty to fifty real cases, which emails the owning team when it breaks. Because the model is swappable, the same suite has to run whenever the model changes too.

“And so regressions will need to cover all of their models.”00:09:03

And the stakes set how strict to be: reads are low risk, reversible writes higher, irreversible actions like sending an email or publishing earnings numbers the highest.

Evidence: transcript 00:09:03, 00:16:05

Lesson 6

Platform and domain teams build it together

No single team solves it. The platform owns the harness, routing and a thin layer of shared tests; domain teams own their answers and their regression cases. Open questions remain, like who owns overlapping data and when sharing is worth the coordination, and he doesn't pretend otherwise.

The reading, for The Focus: the same order for a client. Don't start with the model. Own the infrastructure (a portable harness; frontier or supported open-weight per workload) and own the data and context. Then build the platform on top: who owns each answer, who is asking, evaluations that survive a model swap, and gates by stakes. See the through-line Own the stack.
Read alongsideNetflix engineering leaders on running agents in production, in Anthropic's webinar Scaling AI agent development at Netflix (background, not this talk).

Evidence: transcript 00:17:34, 00:20:28