It’s harder to get a job at Goldman Sachs than it is to get into Harvard.

No, really. This year, the nation’s oldest university accepted 3.6% of its freshman applicants. Per CEO David Solomon, esteemed investment bank Goldman Sachs received “a little over” 300,000 applications for entry-level jobs from new college grads last year—and they hired about 2,500. That’s under a 1% acceptance rate, give or take—a percentage Solomon takes pride in.

“We recruit from a very broad selection of schools and universities all over the world; that’s one of the things that we’ve really tried to expand,” Solomon told Bloomberg’s David Rubinstein in an interview Thursday. “We look for a great work ethic, and majoring in a STEM kind of subject. We have over 10,000 engineers at the firm, so we hire a lot of engineers.”

On the whole, the bank looks for smart, hardworking grads who “believe in excellence, want to win, and have proven that they’ve got grit and determination and an ability to both succeed, but also when they fail to pick themselves up and dust themselves off and keep going,” he asid.

Those are tall orders, which Solomon knows well. Even the chief executive himself was rejected by the bank—twice—before finally making it past the velvet rope in 1999, he revealed.

Armed with a political science degree from Hamilton College, a young Solomon interviewed at “a bunch of banking institutions” in New York City in the early 1980s, and was eventually hired by the Irving Trust Company, a commercial bank, and placed in its training program. “That’s where I started my career and started to learn about finance, but I did not know much about it before that,” Solomon told Bloomberg.

It wasn’t Goldman, but not for lack of trying. “Goldman Sachs was one of 40 or 50 firms that I sent a letter to asking for an interview,” Solomon recalled. “It was one of the 45 that came back and said, no thank you.”

But that wasn’t the end of the road for the banking hopeful. “I also interviewed at Goldman for an analyst position when I’d been at Irving Trust for about a year, and I did well,” he said. “But ultimately I was turned down in the last interview, by a partner. That would have been in 1985. So I got two rejections from Goldman early in my career.”

Yet Solomon stayed the course. He was recruited to the sales desk at erstwhile investment bank Drexel Burnham Lambert, and then moved to (also erstwhile) Bear Stearns. Finally, Goldman came back around in 1999 and recruited Solomon to make a lateral move shortly after the firm went public.

“I was in a relatively senior position at Bear Stearns. And I really wasn’t interested [in Goldman],” Solomon told Rubinstein. “But [the company] kept calling and calling and calling. And he caught me on a day where suddenly I said, you know what? I should really look at this. And so in the summer of 1999, about two months after the IPO, I joined the firm.”

But he never had any aspirations towards joining the C-suite, Solomon said; he simply intended to “learn, grow and do more of the business.” And besides, he said, “there was no way you could come to Goldman Sachs laterally and expect you’d ever be the CEO or a managing partner.”

Staying on top

Yet fortunes are always changing: This year will mark Solomon’s sixth in the corner office. During his tenure, Rubinstein pointed out, Goldman’s stock and market cap have “more or less doubled.”

Indeed, Solomon is attuned to those major barometers. “We’re a public company, and we’re here to create value for our shareholders,” he said. “In the last six years, I think we’ve grown the market cap close to $100 billion… We’ve taken the average revenues of the firm from $34 or $35 billion to over $50 billion today.”

Then there are the more abstract goals. “When you’re stewarding a great organization like this, other things are important and lead to growth and improvement,” he said. “The first is, we are a client business, and if our clients don’t trust the firm [or] the people serving them—if we’re not building trust in long-term relationships and serving them with excellence and distinction—we won’t get those other outputs.”