# Burt

> Burt (YC W26) builds AI teammates for logistics. The San Francisco company helps freight brokers and forwarders deploy AI coworkers that handle repetitive, manual work end-to-end - check calls, document processing, covering loads, invoicing - by reading emails, making calls, logging into portals, and updating the customer's TMS, following each team's own processes. Founded by Bobby Zhong and Kurt Sharma, both of whom grew up in logistics families, Burt started by training specialized models that beat frontier models on cost and latency, then focused those models on freight operations.

- **Founded:** 2026
- **Headquarters:** San Francisco, United States
- **Founders:** Bobby Zhong (Co-Founder & CEO), Kurt Sharma (Co-Founder & CTO)
- **Team size:** 2-3
- **Products:** AI Teammates for Logistics, Specialized model training and deployment
- **Notable:** Accepted into Y Combinator's Winter 2026 batch., Built an in-house post-training and inference stack for specialized models., Delivered a small vision language model ~3x faster (p50) and more accurate than gemini-3-flash for a customer's high-volume workload.

## Products & services

- **AI Teammates for Logistics** — AI coworkers that handle freight workflows end-to-end - shipment tracking and check calls, document processing, chasing PODs and rate confirmations, covering loads, invoice auditing, quoting and load building - by reading emails, making calls, logging into portals, and updating the customer's TMS. They follow each team's specific processes and escalate to a human only when needed.
- **Specialized model training and deployment** — Burt's in-house post-training and inference stack that produces custom, domain-specific models which outperform general closed-source alternatives while being roughly 10x faster and cheaper. In one case study Burt built a small vision language model ~3x faster (p50) and more accurate than gemini-3-flash.

## Achievements

- Accepted into Y Combinator's Winter 2026 batch.
- Built an in-house post-training and inference stack for specialized models.
- Delivered a small vision language model ~3x faster (p50) and more accurate than gemini-3-flash for a customer's high-volume workload.
- Launched twice on YC's Launch YC - first as specialized model training, then as AI teammates for logistics.

## Latest updates

- **2026-03** — Launched on Launch YC as 'Burt - AI Teammates for Logistics.'
- **2026-02** — Launched on Launch YC as 'Burt - Train and Deploy Specialized Models,' with a case study on a VLM beating gemini-3-flash.
- **2026-01** — Raised a $130K seed as part of the Y Combinator Winter 2026 batch.

## Links

- Website: https://burthq.com
- LinkedIn: https://www.linkedin.com/company/train-burt
- YouTube: https://youtu.be/Jj7XKdMwE1Q

---

Profile page: https://yespress.io/burt-yc-w26
Published by YesPress — https://yespress.io
Last updated: 2026-07-31