Math Olympiad Silver, Stanford, Google Brain — Vietnam's Top AI Engineer Left OpenAI

Math Olympiad silver. Stanford. Google Brain. xAI. OpenAI. Phạm Hy Hiếu spent fifteen years reaching the front lines of AI. Seven months later, he took his family home.

Math Olympiad Silver, Stanford, Google Brain — Vietnam's Top AI Engineer Left OpenAI

On February 26, 2026, a Vietnamese engineer posted a thread on X that spread fast through tech circles.

"I am burnt out. All the mental health deteriorating that I used to scoff at is real, miserable, scary, and dangerous."

The post was from Phạm Hy Hiếu. He was 34, and he'd spent seven months as an engineer at OpenAI. He said he was leaving the AI front lines and taking his family back to Vietnam.

He used to hear people talk about mental breakdowns at AI companies and thought they were exaggerating. Now he was the one breaking down.

The math kid from Vietnam

Phạm Hy Hiếu was born in Ho Chi Minh City in 1992 and started collecting math prizes early. In sixth grade, he won gold at the International Primary Mathematics Olympiad. In eleventh grade, he represented Vietnam at the 50th International Mathematical Olympiad in Germany and came home with a silver medal.

That medal got him a scholarship. In 2011, at nineteen, he flew to the United States and enrolled in Stanford's computer science program.

During his junior year, Christopher Manning — one of the biggest names in natural language processing — pulled him onto his research team. It was Hiếu's first taste of AI research. The pure math student changed direction.

He graduated in 2015 with the department's Ben Wegbreit Award for best honors thesis.

Six months of failure

In 2017, Google and Carnegie Mellon launched a joint PhD program that let students conduct research inside Google Brain. Only a handful of spots were available worldwide. Phạm Hy Hiếu got one.

His advisor was Lê Viết Quốc, a Vietnamese-born scientist at Google Brain. The research question: could AI automatically design neural network architectures? Hiếu had an idea — let different AI models share parameters so each wouldn't need to be trained from scratch.

When he pitched it, Quốc said: "That idea is crazy."

But he didn't stop him.

The first six months of experiments produced nothing but failures. Every single one. Hiếu grew anxious and frustrated, convinced he was wasting everyone's time. Quốc was the only person who kept telling him not to quit.

Two breakthroughs changed everything.

First, Quốc arranged a thirty-minute meeting with Geoff Hinton — then Google's VP of AI and one of the founding figures of deep learning, later awarded the 2024 Nobel Prize in Physics. Hinton was nearly impossible to get time with, but Quốc made it happen.

Hinton looked at the experimental results and said: "This is terrible." Then he spent the full thirty minutes picking apart what went wrong and pointing to new directions. For Hiếu, the real lesson was simpler: "Someone at that level was willing to take your bad work seriously."

Second, Quốc brought in colleague Barret Zoph, who joined the team and fixed a long list of bugs in the code. Two months later, positive results finally appeared.

They submitted a paper to a top conference. Rejected. Reviewers called the idea "too crazy" and the writing a mess.

Christmas 2017. Everyone was on holiday. Lê Viết Quốc stayed behind, pulling an all-nighter to rewrite the paper and design new experiments. They resubmitted to ICML, the International Conference on Machine Learning. This time it got in.

The paper was called ENAS — Efficient Neural Architecture Search via Parameter Sharing. It cut the computational cost of neural architecture search by a factor of 1,000. Co-authors included Quoc V. Le and Jeff Dean. By 2020, it had been cited over 4,600 times.

Hiếu finished his PhD at CMU in 2021. He was 29.

Toward the front lines

After his doctorate, Phạm Hy Hiếu stayed at Google Brain as a research scientist. In 2020, Forbes Vietnam named him to its "30 Under 30" list.

That year, he gave optimistic interviews about AI to Vietnamese media. He predicted AI would dominate international math olympiads within a decade. He said Vietnamese talent was rising fast, with more Vietnamese faces showing up at top conferences every year.

Around 2023, he left Google and joined startup Augment Computing, helping it reach a $1 billion valuation.

In August 2024, Elon Musk's xAI recruited him. He worked on attention kernel optimization for the Grok-3 model.

That same August, he moved to OpenAI. From his 2011 departure to the U.S. to this moment, it had taken fourteen years to reach the most important AI lab in the world.

Seven months

He didn't say much about what happened inside OpenAI. He called the experience "once-in-a-lifetime" and his colleagues "the absolute best in every way." He helped build "extremely intelligent entities," and he was proud of that.

But the cost, he said, was too high.

"I cannot believe I would say this one day, but I am burnt out. All the mental health deteriorating that I used to scoff at is real, miserable, scary, and dangerous."

He wrote something else too: "Today, I finally feel the existential threat that AI is posing."

A man who spent his entire career building AI says he finally feels threatened by it. That sentence carries its own weight.

He's not the only one

Phạm Hy Hiếu is not an isolated case. Around the same time, Mrinank Sharma — head of safety research at Anthropic, the company behind Claude — also resigned, saying the world faces "multiple intertwined crises."

Nathan Lambert, a researcher at the Allen Institute for AI, has described a culture of self-imposed pressure at companies like OpenAI and Anthropic. Nobody tells you to work overtime. But when everyone around you is pushing to the limit, not pushing feels like falling behind. Some have compared the intensity to China's 996 work culture.

In the AI arms race, every lab is sprinting. Models get stronger each generation. Release cycles compress. The engineers are fuel fed into an accelerator.

Going home

Phạm Hy Hiếu is back in Vietnam with his family. He says he wants to rest, get healthy, and try something different.

As early as April 2025, while still at xAI, Vietnam National University in Ho Chi Minh City had appointed him as a visiting professor — one of the first sixteen scientists selected. His ties to Vietnam never broke.

Fifteen years ago, a Vietnamese teenager left for America with a math olympiad medal.

Now he's come home.

` })