# Yikang Shen: CEO of Learning Machine

Left xAI in 2026, where he was introduced as Pre-Training Lead; earlier a staff research scientist at IBM.

Researched with Autumn (https://www.autumn.ai) from public records. Updated 2026-11-16. 3 sources. Company file: https://www.autumn.ai/resources/stealth-startups/learning-machine

## How Yikang is publicly tied to Learning Machine

His X profile reads "CEO at Learning Machine. ex MTS @xAI. ex Staff RS @IBM. PhD @Mila." [1](https://x.com/Yikang_Shen)

## Career path

- 2017–2022: Mila, PhD student [1](https://www.alphaxiv.org/@yikang-shen)
- 2022–2025: IBM, staff research scientist [1](https://www.alphaxiv.org/@yikang-shen) [2](https://x.com/Yikang_Shen)
- 2025–2026: xAI, member of technical staff; introduced as Pre-Training Lead at a December 2025 talk [1](https://www.alphaxiv.org/@yikang-shen) [2](https://joindelta.org/talks/yikang-shen)
- 2026–: Learning Machine, CEO [1](https://x.com/Yikang_Shen) [2](https://www.alphaxiv.org/@yikang-shen)

## Education

- PhD, Mila [1](https://x.com/Yikang_Shen) [2](https://www.alphaxiv.org/@yikang-shen)

## Notable

- His work on linear attention, including Gated Linear Attention and DeltaNet, was the subject of a research talk at NeurIPS in December 2025. [1](https://joindelta.org/talks/yikang-shen)
- He is a coauthor of Long Range Arena, a benchmark for efficient transformers on long sequences, and of Ordered Neurons. [1](https://www.alphaxiv.org/@yikang-shen)

## Links

- LinkedIn: https://www.linkedin.com/in/yikang-shen-5094b36a
- X: https://x.com/Yikang_Shen

## Learning Machine

AI lab from a former xAI researcher, building foundation models that keep learning after deployment. Still in stealth. https://www.autumn.ai/resources/stealth-startups/learning-machine

Corrections and removal requests: founders@autumnai.com
