Sentient Foundation Names Julie Belião as VP of Product

Sentient Foundation’s four open-source projects have collected more than 12,500 GitHub stars as the San Francisco nonprofit AI research group named Julie Belião vice president of product.

Belião joins from Mozilla.ai, where she held the same title, after earlier roles at Defined.ai, Unbabel and Google. She will direct product development across Sentient’s frontier research programs, turning open-source work into tools for developers and enterprises building self-improving AI systems.

What will Belião build?

Her remit emphasizes recursive self-improvement and model provenance. The stated goal is to turn frontier research into practical products that developers and organizations can build on while making AI models more transparent and trustworthy.

Sentient’s portfolio spans agentic AI, open search and model verification: ROMA, a recursive open meta-agent framework; OpenDeepSearch; OML model fingerprinting; and EvoSkill. Sentient says thousands of developers worldwide have adopted the work and that researchers at Google, Microsoft, Meta and leading universities have cited its output.

What does Belião say?

“The AI industry has become very good at building impressive technology. The harder question is how we turn that technology into something genuinely useful, trustworthy, and accessible.”

“What excites me about Sentient is the opportunity to tackle fundamental challenges like recursive self-improvement and turn frontier research into practical tools that developers and organizations can actually build on,” said Julie Belião, VP of Product at Sentient Foundation.

“Julie brings a unique combination of experience across academia, major technology companies, and startups, along with a proven ability to build products from the ground up.”

“As Sentient focuses on translating frontier research into practical products, her expertise will be invaluable. We’re thrilled to welcome her to the team,” said Sachi Kamiya, Director of Growth and Venture at Sentient Foundation.

What is driving the commercial push?

Interest in recursive self-improvement is intensifying. The field drew more than $2 billion in venture funding over the past year, according to Sentient’s account, while the foundation said Google DeepMind is already exploring the approach in production.

Model provenance remains an emerging field without an agreed verification standard, Sentient said, strengthening the case for independent approaches to transparency and trust. The foundation also seeks to position itself as a neutral party because no single company can credibly serve as both builder and auditor.

In September, Nvidia agreed to acquire Hugging Face for approximately $11.9 billion, based on FinanceFeeds’ reading of an Nvidia 8-K filing. The target platform served more than 18 million developers and 3 million models, giving Nvidia’s proposed combination exposure to a major open-source AI distribution channel.

Concern about unchecked autonomous agents is growing alongside that investment. Elon Musk warned that such systems could one day target financial systems, adding a security context to Sentient’s work on transparent, self-improving agents.

What does she bring?

At Mozilla.ai, Belião led open-source ecosystem strategy and community growth across 61 repositories. Sentient credits her with helping establish Mozilla as a recognized contributor to the agentic AI developer ecosystem, a record that matters as open-source tooling becomes a commercial battleground.

She spent a decade working in multilingual AI on speech synthesis, linguistic resources and natural-language-processing data for Google Assistant, Translate and YouTube. Her PhD in computer science from Université Paris Nanterre examined how sentence structure relates to speech patterns in spontaneous conversation.

Belião also co-authored “Rhapsodie: a Prosodic-Syntactic Treebank for Spoken French” in 2014, a shared dataset for spoken-language research. In 2025, she joined Mozilla–EleutherAI collaborators on “Towards Best Practices for Open Datasets for LLM Training,” which recommends how open training data should be sourced, documented and shared.

Can research become a product business?

Sentient says its mission is to keep advanced AI within everyone’s reach by keeping the most powerful systems open source and narrowing the distance between closed frontier labs and the wider AI ecosystem. It also casts itself as a bridge between open-source AI and trust infrastructure, where commercial and audit interests remain separate.

With more than $2 billion in recent venture backing and Nvidia’s $11.9 billion move to acquire Hugging Face, Belião’s mandate now tests whether independent research can turn recursive agents and provenance tools into widely adopted products.