# Vault: Making Unstructured Data Work for Investing — Aaron Linsky, Sam Green — session 2026-10-13T13:00:00.000Z → 2026-10-13T13:18:00.000Z

_53 transcript lines · 32 slides_

## Transcript

So I'm super excited to be part of this with you all today. And now let's hop into what's gonna happen. First of all, welcome to AI Engineer New York twenty twenty six, where AI engineering meets Wall Street. This is the conference for engineers and technical leaders building production AI across banking, hedge funds, fintech, insurance, and asset management. Over the next two days, we will hear from companies and leaders, sparking discussion not only on models, Agents, infra, compute, dev tools, and evals. But then, how they weave and shape the financial industry, impacting hedge funds, investors, banks, insurance, fintech, and security. Of course, as you all can tell, there's a lot going on. Putting together a gathering like this requires a huge amount of effort, And it could not be done without our sponsors. But before our sponsors, let's all just give a huge shout out to the AIE team. What they do is incredible. Now let's give some flowers to the sponsors. First up, the presenting sponsor, Arise. Let's give it up for Arise. Yeah, Arise. Next up, the platinum sponsors Oracle, Sonar, Work OS, and Excel. Let's give it up for them. Then we have our gold sponsors. If I had enough time, I'd rattle all of them off. And finally, the silver sponsors. But wait. There's more the supporting sponsors. Awesome. So we are going to hop into, the first keynote of the day. We have two amazing folks from Bridgewater who are in the weeds of production AI every single day. So let's give it up for them. Welcome, guys. Coming up next, we're delighted to welcome two Speakers from Anthropic. Please put your hands together for Belinda Neal, Managing Director of Financial Services, and Kian Shahab, Applied AI Architect. Just waiting for our slides. There we go. Good morning, everyone. I'm Sam Green, team lead for Applied AI at Bridgewater Associates, and this is my colleague, Aaron Linsky. And today, we're here to talk to you about the Vault, the data platform that we've built at Bridgewater to turn unstructured data into a first class asset in our investment process. And as part of the talk this morning, we'd like to share three lessons that we've learned building the vault. The first is that humans and agents have very different needs from a data Platform. And so you have to build for those different needs separately. The second is that the greatest impact occurs when a single tool can meet both of these sets of needs at the same time. And the third is that the real prize is when you can build the foundations needed to make agents a proactive participant in your investment process or whatever you're doing with your with your your tools, instead of a reactive, reactive tool for your humans to use. Before I dive deeper into talking about the vault, I wanted to start with some some background about Bridgewater. Over the last fifty years, Bridgewater has been in the constant pursuit of the world's best understanding of how markets and economies work, and turning that into a systematic encoding that we can use to stress test our logic over time. And what originally began as a set of if then rules on a yellow legal pad evolved into a continuous compounding process with humans at the center, constantly looking For the next thing that's changing in markets. And more recently, we've been working on building Aya, the artificial investment associate, which puts a machine at the center of this continuous process. Across all these phases of evolution, we've had a constant philosophy in mind, WHICH IS THAT WE WANT TO BE CONSTANTLY LOOKING AT How THE World IS CHANGING, AND WHAT'S GOING ON IN THE World, NOTICING WHAT DYNAMICS ARE NOT YET ENCODED IN OUR UNDERSTANDING, AND USE THAT LEARNING TO IMPROVE AND Pound for the future. For most of Bridgewater's history, that meant building and overseeing time series models on structured data. And that structured data needed to be carefully built so that it could be safe for use in our research process and aligned to our internal ontology and world model so that we could get great leverage out of using the the data that we had built. With the emergence of language models, we had a huge new opportunity to engage with a with a much wider world of data. You know, news and research, transcripts, policy documents, and all of the internal conversation that was previously just in the ether. So the internal memos and meeting transcripts and debate happening inside the firm. We quickly realized that all of that data needed the exact same treatment to be fit for purpose. It needed to be carefully cleaned, governed, and made ready for use in both point in time research questions today and also our systematic study over time. And when we looked out at the landscape of data vendors and data platform providers, there was nothing that provided us with all of the control that we needed to build and maintain our data advantage. And so we built The Vault, which is our repository of external and internal unstructured data, and the platform that indexes, organizes, and makes all of it ready for use by our human investors and agents. Today, the ball Has more than a 100,000,000 documents in it with thousands more added every day, minute by minute, from a 135 external sources, and, again, growing all the time, as well as all of the data from inside of Bridgewater. And crucially, all that data is not just dumped on our investors' desks. It is carefully aligned to our internal internal ontology, with, enrichments like the countries, concepts, markets, and most importantly, time period discussed in the Document. To bring this to life, I want to start with a quick motivating example. So what I did here is I asked a fairly basic representative investment question to an enterprise chat client. I asked, how is AI infrastructure investment impacting the Korean won, and do market participants expect this to continue? And as as you're all, I'm sure, familiar with, the agent goes to the web and looks for resources they can draw on, because I've explicitly asked it. To go find, other people's perspectives on the question. And you can see it zooms in, and it finds 15 relevant looking resources to draw on, before it gives me back its answer. AND WHEN I ZOOMED IN ON THE ANSWER IT GABE BACK, A COUPLE OF THINGS STOOD OUT TO ME. THE FIRST IS THAT EVEN THOUGH IT FOUND 15 RELEVANT LOOKING SOURCES, IT ONLY PICKED ONE OR TWO to answer to answer my question. And the second is down at the end of this response where the agent, it's SELF acknowledges that it hasn't answered my question. It didn't find enough documents on the open web to establish a market consensus and meet the meet the parameters that I laid out. When I double clicked even further, the problem got even worse. ONE OF THE NOTES THAT IT PRIMARILY DREW ON TO ANSWER MY QUESTION WAS PUBLISHED WAY BACK IN July 2026, EVEN THOUGH I ASKED THIS QUESTION AND CAPTURED THIS ANSWER Just AT THE END OF LAST WEEK IN EARLY OCTOBER. In the age of AI and in investing, three months is a massive amount of time. It's not even close to good enough to draw on this resource to answer my question. When I asked this exact same query attached to the vault, SOMETHING TOTALLY DIFFERENT HAPPENS. THE FIRST IS THAT THE CONNECTION TO THE VALT HELPS TO PROMPT THE AGENT THAT IT SHOULD LOOK FOR A SPECIFIC PERIOD OF TIME WHEN ANSWERING MY QUESTION, SO IT ZOOMS IN ON October 1 TO October 8. IT FINDS A MUCH WIDER ARRAY Resources, 52 candidate documents that are relevant to the query over this specific period of time because of the vault's much more significant corpus of content that's relevant to investing questions. And when it ultimately answers my question, it draws on seven instead of only one or two, and gives me back a rich answer that the agent is more confident is going to address my question. And so for us, this is the the big lesson of... From building the vault, to kind of motivate. The talk today, which is that it takes great data and the right connectors to make agents impactful. And I'm going to hand it over to Aaron now to talk about some of the product engineering takeaways from building the Vault. Thanks, Sam. So hi, I'm Aaron. I run Product Engineering at Bridgewater. I'm going to take us through a few of our learnings from building the Vault over the past few years. I'm going to start with one that might seem obvious, but I still think it's worth saying, which is search is hard. But It gets even harder if we're designing for two different consumers. And in this case, we're designing for humans and agents. I'm gonna go over the difference here, starting with us. So humans need precision and a high signal to noise ratio because we can only read so much. Even a fast reader manages about 300 words a minute. And personally, I'm much slower. Agents are a different kind of reader. They're far less sensitive to precision. If the data is there, they'll find it. And basically, Anything you hand to them will get attended to. They scale super quickly, and they come and they consume about, 100 times faster, 30,000 words per minute. And so we need to build Vault for both of these. And when we're thinking about the dimensions that matter, there are these five on the board recall, precision, filtering, entitlements, and scale. Some of them matter to people, some matter more to agents, many to both. I'm gonna go deep on three of these, precision for people, scale for agents, and then filtering for both. So precision. Once we'd filled the vault, our... Which was a heroic effort by our data and legal teams, recall was no longer a pressing issue. However, the number one complaint was Vault Search is bad. As a Product Owner, it's painful to hear that. But my response was, let's benchmark it. And that's what we did. We collected questions and found the best documents by hand, including ones that looked helpful, but weren't. In total, it was about 200 examples, and it took us less than a week. It was not fun or very glamorous, but it was very helpful. And after we ran it, it was very clear Search was bad. The key metric we were looking at is precision at 10, which is, of the top 10 results, how many are the right documents? And we were hitting about 60%. Percent. So the people were right. It was bad. But now we had a way to measure all of the improvements. And here on the board are five of them. And by the time we got done with the agentic search, we were up in the 90%, 95% recall at 10 or precision at 10. All right, so that was precision. Now on to filtering. Like anyone, we want the most relevant documents. But when we looked at the production traces, what we were seeing is That almost always, there was some condition, either as of a certain date, or from a certain source, or about a specific Asset. And all of that requires advanced filtering. And that's where we ran into our first major performance issue. So I'm going to quickly go through this of a little bit technical here. There's classically two ways to filter a vector database. Post Filtering, we'll use that vector index first, and then filter the top results by metadata. It's fast, but recall could be pretty low, because the globally relevant results might not match your filters. On the other side, we have pre filtering, where we'll filter first and then calculate distances for every one of those matches. The advantage there is that the recall is very good, but it can be very slow. And we needed something in between. So that is what led us to try, and then ultimately switch to TurboPuffer. Their approach, which they call Native Filtering, looks only in the nearest Clusters that contain Matches, and then scores just those. What you end up getting is something that's fast, like post filtering, but with close to the same recall as pre filtering. And lastly, I'm going to talk about scale. Human scale for us is pretty pedestrian 100 queries an hour. We have 300 users. They're going to search, you know, every so often. It takes our users minutes to read the Docs that they're receiving. Agents, on the other hand, are a few orders of magnitude. Beyond that. Easily 100 plus queries a second, and Agent swarms can hit thousands easily. And they obviously will read, as I said earlier, will read docs in seconds. So we needed something that could scale as well. And again, TurboPuffer handled that easily too. So we were covering all five dimensions. I didn't talk about entitlements, but I'll be talking about that later today. But we covered it all. It unlocked new workflows, but it didn't change how our researchers worked. They still Weren't coming to Vault every morning. And that's my second lesson that I wanna share here is that no matter how amazing the tools are, we can't just drop them on our investors' desks and expect them to be used. We need to migrate their old habits into a new form. And so we looked at what they were already doing. Many of them were living in email. The research came straight from the brokers. They had subscribed lists. They had rules that filtered everything into folders. It worked. It was familiar, and they knew that they weren't missing anything. So we had to build all of these Features into the Vault. So that's exactly what we did everything's searchable. The subscribed lists became Saved Searches, and Folders became Saved Collections. It's the same Habits, just a new Home. And once we had them there, we can open up new Habits. So they can ask questions across all of the data, as Sam showed. They can follow dynamic feeds by topic, and they can hand those Collections over to their Agents to take further action. And that's when we clearly had Product Market fit. And once we had that fit, the users wanted it everywhere, so we put it in Chat. We made a phone app. It's in Slack in their IDE. And of course, we availed all of that to the Agents as well, including a CLI, because Agents love CLIs. And so the main question here is, okay, so what does this enable? Sam's going to talk about what we're building right now. But first and foremost, it's helping us to move around That research circle faster. With our Pocket Analyst, which is our first AI analyst, it can do 100... It can do hours... It can do hours of research in just a few minutes for a few dollars, and that's relying on the Vault. And then we're expanding that now with the Researcher, which is the next generation. It's the same idea, stretched across days or weeks. We're also able to train Custom Models to more intelligently reason over the Vault's data, and use those together with other Foundation models to feed systematic decision making processes. The Forecaster is our first one, which produces AI generated macro relevant predictions. And then we also have built investment signals from the Vault, turning unstructured data into structured, tradable strategies. Now I will hand it to Sam to take us through Vault's latest capability. Thanks. So like I alluded to at the beginning of the talk, the real prize is When you can take agents from a reactive resource that can be ready to answer a human's question, and instead make them proactive participants, helping to steer your high value professionals towards the most important areas for their attention. And the way that we're solving that problem is in an experience we're calling the investor workbench, where the agent is always on, tracking what's important to that investor, and presenting them with the areas that are most in need of their attention. It looks like an experience like this that relies on the vault, where a topic is proposed and several pieces of information are presented to the investor to decide whether they should dig deeper or ignore that topic entirely. You can see here the card that an investor sees in order to decide whether to dig in deeper into the same AI impact on currencies question that we looked at earlier in the talk. And the investor receives everything they need. An agentic summary if they want a quick scan across the agent's Findings. They also get the exact evidence that the agent was drawing on, so they can assess for themselves what was leading to this recommendation. And if they choose to dig deeper, they have a direct connection to the research output from the analytical tools that Aaron was just describing. And when they dig deeper alongside the agent to look at the dynamics driving this question, the learnings are automatically captured. The results of their collaboration are automatically captured back.

## Slides

### 00:01:44

# AI Engineer NEW YORK

## PRESENTED BY
- Arize

## PLATINUM SPONSORS
- Oracle AI Database
- Sonar (Makers of SonarQube)
- WorkOS
- exa

[Logo for Arize]
[Logo for Oracle AI Database]
[Logo for Sonar]
[Logo for WorkOS]
[Logo for exa]

### 00:02:52

### Managing Director, Financial Services
ANTHROPIC

### Applied AI Architect
ANTHROPIC

[Stylized image of the New York City skyline with the Statue of Liberty]

### 00:03:22

### Sam Green · Team Lead, Applied AI
### Aaron Linsky · Head of Product Engineering, Bridgewater Associates

## The Vault turns unstructured data into a first-class asset.

### 00:03:50

# The Vault turns unstructured data into a first-class asset.

1. Humans and agents have different data platform needs
2. The greatest impact occurs when these different needs can be met by a single tool
3. The real prize is to shift agents from reactive to proactive

### 00:04:20

# FIFTY YEARS OF COMPOUNDING
## Pursuing the best understanding of how markets and economies work.

-   **The 1980 bond system** (handwritten notes on

### 00:04:51

# FIFTY YEARS OF COMPOUNDING
## Pursuing the best understanding of how markets and economies work.

### The 1980 bond system

### Our Learning Process
###

### 00:05:19

# From tabular data to unstructured data.

### BEFORE 2023 • TABULAR DATA
Time series models on structured data
Built to align to our ontology and world model

[Diagram

### 00:05:49

# From tabular data to unstructured data.

## BEFORE 2023 - TABULAR DATA
### Time series models on structured data
Built to align to our ontology and world model

## SINCE 2023 - UNSTRUCTURED DATA
### Language models with unstructured content
- External: News, research, transcripts, policy memos
- Internal: Live debate, chat conversations, memos

[Diagram showing a transition from "tabular data" (before 2023) to "unstructured data" (since 2023), represented by two boxes with an arrow in between.]

### 00:06:17

# So we built the Vault.
A repository of external and internal unstructured data, and the platform that indexes, organizes and secures it.

~100M
documents, with thousands added every day

135
external sources from around the world

[3D diagram showing three stacked translucent cubes receding into the distance, with an arrow labeled "TIME" pointing along the axis of recession. The front cube is labeled "Concept", "Country", and "Market" on its visible faces.]

### 00:06:51

# So we built the Vault.
A repository of external and internal unstructured data, and the platform that indexes, organizes and secures it.
- ~100M documents, with thousands added every day
- 135 external sources from around the world

[Diagram showing three stacked, receding cubes, with the front cube labeled "Concept", "Country", and "Market" on its faces, and an arrow pointing right labeled "TIME"]

### 00:07:19

### SAME QUESTION, SAME LLM
# Standard tool: out of date, under-sourced

> **User Query:** How is AI infrastructure investment impacting the Korean Won? Do market participants

### 00:07:48

### SAME QUESTION, SAME LLM
# Standard tool: out of date, under-sourced
1. Searches the open internet and reads several sites.
2. Builds its answer on one

### 00:08:26

### SAME QUESTION, SAME LLM
# Standard tool: out of date, under-sourced

1.  Searches the open internet and reads several sites.
2.  Builds its answer

### 00:08:47

## SAME QUESTION, SAME LLM
# The Vault: Searching with Time as Part of the Query

[Screenshot of an AI assistant interface showing a user query: "How is AI infrastructure investment

### 00:09:24

## SAME QUESTION, SAME LLM
# The Vault: Searching with Time as Part of the Query

1.  Reasons that it needs recent data: October 1 to 8.
2

### 00:09:50

# Product Engineering Takeaways
## Aaron Linsky - Head of Product Engineering

### 00:10:22

# Two very different consumers

## HUMANS
- Sensitive to Precision
- High signal-to-noise ratio
- Small scale
- Slow to consume
- ~300 words per minute

## AGENTS

### 00:10:53

# Two very different consumers

### HUMANS
- Sensitive to **Precision**
- High signal-to-noise ratio
- Small scale
- Slow to consume
- ~30

### 00:11:21

# Different needs
## FIVE DIMENSIONS
- Recall
- **Precision**
- Filtering
- Entitlements
- Scale

### 00:11:49

### PRECISION • FOR PEOPLE
# #1 Complaint: “Vault search is bad”

## RESPONSE
### Let's benchmark it

- Collect questions, manually find the best docs, including marking ones that were bad
- About **200** in total

### Search WAS bad
Enabled us to measure our improvements:

### 00:12:19

### Different needs

# #1 Complaint: “Vault search is bad”

## Response: Let's benchmark it

- Collect questions, manually find the best docs, including marking ones that were bad
- About **200** in total

### Search WAS bad

Enabled us to measure our improvements:

- Fixed Doc ingest/parsing issues
- Reciprocal Rank Fusion weights
- Reranker
- Query expansion
- Agentic Search

### 00:12:51

## FILTERING - FOR BOTH

### We want the most relevant documents

- as of July 16, 2026
- U.S. Federal Reserve
- KRW

### 00:13:20

### Different needs
#### FILTERING • FOR BOTH
# Postfiltering v Prefiltering

### POSTFILTERING
- ~10% recall **Fast**
- Search the whole index,

### 00:13:48

### Different needs
## SCALE • FOR AGENTS
# Need to serve at Agent Scale

### HUMAN SCALE
- 100 Queries per Hour
- Minutes to read Docs

### 00:14:19

### Different needs
## Need to serve at Agent Scale
### SCALE • FOR AGENTS

#### HUMAN SCALE
- 100 Queries per Hour
- Minutes to read Docs

#### AGENT SCALE
- 100+ Queries per Second
- Docs read in seconds

### 00:14:52

# Vault: When Fifty Years of Structured Data Isn't Enough

## Vault covered all five dimensions
- Recall
- Precision
- Filtering
- Entitlements
- Scale

It unlocked

### 00:15:24

### Two consumers, one tool

# Same habits, new home -> new habits

**Port**

- **IN EMAIL**
  - Research received by email
  - Subscribed lists

### 00:15:49

### Two consumers, one tool
# Wherever the work happens.
Ramps to expand their tools, from the browser to the agents they direct.

## PEOPLE
- Browser
- Chat
- Phone
- Slack
- IDE

## AGENTS
- API and Python client
- Command-line tool and skill for coding agents
- MCP server

### 00:16:21

# What agent access makes possible.

## HUMAN ACCELERANTS

- **Pocket Analyst**
  Our first AI analyst. Hours of research in minutes, using the Vault.
- **Researcher**
  The next generation of our Pocket Analyst capable of performing long time horizon financial research projects around our entire research circle.

[Icon of a briefcase next to "Pocket Analyst"]
[Icon of a telescope next to "Researcher"]

### 00:16:47

# What agent access makes possible.

## HUMAN ACCELERANTS
- **Pocket Analyst**
  Our first AI analyst. Hours of research in minutes, using the Vault.
- **Researcher**
  The next generation of our Pocket Analyst capable of performing long time horizon financial research projects around our entire research circle.

## AGENT SANDBOXES
- **Custom RL models**
  Models we train with reinforcement

### 00:17:22

# Building the Investor Workbench
### Reactive to proactive

- **REACTIVE**
  - Answers when asked
- **PROACTIVE**
  - Steers attention

[Flow diagram showing a transition from "Reactive" to "Proactive"]

### 00:17:53

# Summary and evidence, ready for action.

### Screenshot of a research document titled "AI impact on currencies"
### Screenshot of an "Activity Feed" showing a list of research and news articles