---
session_id: "1373462"
title: "How to build agents over your own data"
phase: live
last_checked: 2026-11-12T13:24:21-08:00
render: 20261112-132421-c4d6ee
---

# Links: How to build agents over your own data

## Speaker

**Aamir Shakir**, Founder and CEO, Mixedbread

The roster published no links for him, and no LinkedIn, X or GitHub profile could
be confirmed as his (`github.com/mixedbread-ai` is the company's).

## Company

**[Mixedbread](https://www.mixedbread.com/)**: "Mixedbread turns PDFs, decks,
videos, code, Slack, and Drive into evidence your agent can act on, with fewer
tokens." (from the site)

## Resources

- **Product** · [Mixedbread Connectors](https://www.mixedbread.com/docs/connectors).
  Slack, Google Drive, Gmail, Linear, Notion, GitHub and more, synced into one
  index instead of one MCP server per source.
- **Blog** · [Introducing Mixedbread Search](https://www.mixedbread.com/blog/mixedbread-search)
  (2025-10-01). The multimodal, multilingual search API that the single index is
  built on.
- **Blog** · [Closing the Oracle Gap for Your Agents](https://www.mixedbread.com/blog/closing-gap)
  (2026-03-24). Mixedbread Search v3 for agentic retrieval, measured against a
  browsing oracle.
- **Blog** · [Beyond the Limit: Introduce Mixedbread Wholembed v3](https://www.mixedbread.com/blog/wholembed-v3)
  (2026-03-12). The retrieval model behind the scaling-laws chart and the
  "ten-point change" over the previous model.
- **Blog** · [The Hidden Ceiling: How OCR Quality Limits RAG Performance](https://www.mixedbread.com/blog/the-hidden-ceiling)
  (2025-05-14). Why a build-it-yourself pipeline pays for OCR and parsing, and
  why a multimodal index beats even perfect text.

## Gaps

- No personal profile could be verified. The roster lists none.
- The bank case study, which the talk says will be published in two days (around
  2026-11-13), is not out yet. The next recheck should look for it.
- No official recording or slides found yet. The dossier rechecks every 6 hours
  during the event.