Skip to content
Jev Hub

[ Learn ]

Who founded TypeSafe AI? The team behind Jev

Who founded TypeSafe AI, maker of Jev: Diogo Almeida, Erik Gafni and Sasha Sheng. What the company and press reports say about each. Checked 1 November 2026.

Published By Jev Hub

[ Sponsors ]

These sponsors help keep Jev Hub updated. Want yours up here?

Back to Learn
[ TLDR ]
  • TypeSafe AI, the San Francisco company that makes Jev, was founded by Diogo Almeida, its chief executive, together with co-founders Erik Gafni and Sasha Sheng.
  • Almeida wrote the Jev launch post on 15 September 2026, and TypeSafe's documentation calls him a cofounder.
  • The co-founders' names come from reports of the seed round by SiliconANGLE, Dealroom and Value Add VC; we could not open the company's own press release, and we found no title or background for Gafni or Sheng in a source we trust, so we give none.
  • What the company says about Almeida's earlier work at OpenAI is its own account, which we have not independently confirmed.

The founders at a glance

FeatureWhat is knownWhere it comes from
Diogo AlmeidaFounder and chief executive of TypeSafe AI; the documentation calls him a cofounder. Says he helped build instruction-following methods at OpenAI, and the documentation says he co-invented RLHFTypeSafe's launch post and AI primer (his and the company's own claims)
Erik GafniNamed as a co-founder. No title or background that we could confirmSiliconANGLE, Dealroom and Value Add VC, reporting the seed round
Sasha ShengNamed as a co-founder. No title or background that we could confirmSiliconANGLE, Dealroom and Value Add VC, reporting the seed round
The companyTypeSafe AI, San Francisco, maker of Jev; came out of stealth on 15 September 2026TypeSafe's own pages
On XAlmeida posts as @CompleteSkeptic; the company as @typesafeaiTheir own accounts

Read on 1 November 2026. Statements about Almeida's earlier work are TypeSafe's and his own claims, labelled as such, and this site has not independently confirmed them. The co-founders' names come from press reports of the seed round, not from a page we could open at the company. We give no detail we could not source.

Who founded TypeSafe AI?

TypeSafe AI was founded by Diogo Almeida, its chief executive, with co-founders Erik Gafni and Sasha Sheng. Three outlets that reported the company's $40 million seed round, SiliconANGLE, Dealroom and Value Add VC, name all three. SiliconANGLE adds that the two co-founders also have extensive AI experience, without giving more detail.

TypeSafe's own pages that we could read name Almeida: his byline is on the launch post, which calls him founder, and the documentation calls him a cofounder. We could not open the company's press release, so the two co-founders' names are as reported rather than confirmed by us. We found no title or background for them in a source we trust, so we give none.

What does the company say about Diogo Almeida?

In the launch post he says that after two years in stealth the company released its first System One model, Jev, on 15 September 2026. He says that at OpenAI he helped build the methods that made language models useful at following instructions and talking with people, and that this work ended up as the research behind ChatGPT. TypeSafe's AI primer goes further: it says RLHF was used to train InstructGPT and ChatGPT and was co-invented by him, and links to his Google Scholar page.

Both are the company's own statements about its founder. Press coverage calls him a former OpenAI researcher, which fits, but we have not checked the details against OpenAI.

01Instructionfollowingat OpenAI,his account02Two yearsin stealthTypeSafe AIbuilds Jev03Jevlaunches15 Sep 2026early access04Runtimetalkpublished6 Oct 2026
The path to Jev, as the launch post tells it. From TypeSafe's launch post of 15 September 2026 and Modal's Runtime 2026 talk. The first step is the founder's own account.

What do we know about the co-founders?

Less, and we say so. Erik Gafni and Sasha Sheng appear by name in the seed-round coverage as co-founders, and nowhere else that we could verify. Some profile and wiki sites list titles and earlier employers for them, but those sites do not cite the company or a primary source, so we do not repeat them.

If TypeSafe publishes a team page or a press release naming their roles, this article will be updated with it and dated.

Where can you hear from the team?

Two long recordings of Almeida are listed on this site: a talk at Modal's Runtime 2026 conference, published on 6 October 2026 and about 13 minutes long, and a long interview with the Latent Space podcast, published on 21 September 2026. He posts on X as @CompleteSkeptic and the company posts as @typesafeai. We found no recordings or accounts that TypeSafe itself points to for the co-founders.

A post by a founder is a company statement, so we treat it as a claim until a primary page, such as the documentation or blog, confirms it.

Our read

People check who is behind a model before they build on it. Here the answer is three named founders, a chief executive with a stated research background at OpenAI, and a company that has raised a large amount in a short time. Only one of the three has material of their own that we can read, so most of what is known is about him.

That is the company's story and the press's, and it should not stand in for testing. Judge Jev on your own data, and treat what any founder says as a claim to check.

Common questions

Sources

Written by Jev Hub from the sources below and checked against them on . Where a figure is a company's own claim, the article says so.

Related content