The Operator - Adam Gramling on leading Global Accounts and Programs at Scale AI, building client delivery playbooks, and launching SR² Ventures
How Adam went from Chief of Staff to running Scale's largest accounts - and what he learned about AI companies along the way.
Hi, I’m Jan. After years as a Chief of Staff and Operations Lead at companies like Airbnb, I started The Operator to give a voice to the people leading, growing, and shaping high-growth tech companies. Every few weeks, I share stories from Operators who have helped build some of the most exciting organizations in tech.
Setting the Stage
Scale AI is the dominant player in human data labeling for artificial intelligence. The company powers the training of frontier AI models (think OpenAI, Anthropic, Meta) by providing the human feedback and data annotation that make LLMs work.
Founded by Alexandr Wang and Lucy Guo in 2016, Scale AI has grown into one of AI’s most critical infrastructure companies. In 2025, Meta bought a 49% stake in Scale AI for $14B and integrated some of their key technologies and teams.
For this conversation, I sat with Adam Gramling, who joined Scale AI in 2022 and spent three and a half years running some of the company’s largest enterprise accounts. Adam initially joined as an Engagement Manager and eventually rose to start and lead both Technical Account Management and Program Management for GenAI. Today, Adam runs SR² Ventures, his own AI-focused investment fund and advisory practice.
Adam’s background spans management consulting, corporate strategy, and a Chief of Staff role in a real estate development company. Today, his portfolio at SR² includes investments in YC companies including several robotics and aerospace startups.
As a fellow generalist navigating the AI space, I found Adam’s perspective on client delivery, onboarding frameworks, and pattern-matching from Operator to Investor invaluable. I hope you enjoy this issue as much as I did.
From Engineering to Business
Adam’s path to land at Scale AI wasn’t really linear. Following the advice of some Silicon Valley executives he had interviewed, Adam studied industrial and systems engineering at USC. Post-graduation, he went into Management Consulting to work across data analytics projects. When an unexpected opportunity came up, Adam became Chief of Staff at a leading real estate development company:
“I moved out to New York, wore a suit and tie, and worked on Wall Street. That was a very interesting role, quite different from working on data projects at tech companies in San Francisco. It was fast paced and I had to get up to speed, research key initiatives, and deliver excellent work on very tight timelines.”
When the macro environment shifted, he moved to NetApp, a data infrastructure company, to work in corporate strategy. Adam deepened his understanding of data and AI. He wasn’t looking to leave when Scale AI reached out.
Joining Scale AI: Data as the backbone of the AI-age
Interviewing with Scale AI, Adam wanted to ensure to find a role that would match his previous background and ambitions. He ended up joining as an Engagement Manager for Scale’s largest customer:
“I absolutely loved that role. You were essentially a mini-CEO running everything for that project: your revenue, cost, dealing with client management, building out new product features. I had a lot of fun doing that.”
What Adam found at Scale AI was a company moving at breakneck speed in a market that was about to explode. When ChatGPT went viral in November 2022, the demand for training data skyrocketed. Scale AI was poised for growth and needed to succeed quickly in a competitive environment.
Speed as a Core Value
Asked for his biggest learnings from Scale AI, Adam’s answer was immediate: speed.
“[The time at Scale AI] really taught me to just get things done. No more ‘hey, let’s set a chat next week to talk about it.’ No, let’s go solve this right now. We’ll set an hour, we will knock it out and be done with it. You’re moving so fast, you just can’t delay things.”
The second lesson was about transparency:
“Be very crisp. Know your numbers, know your data, know why things are working. And don’t try to BS people, because if you do, they will tear you apart.”
The Four-Week Onboarding Framework
As Scale AI was in hyper-growth mode, Adam and his peers needed to onboard new team members at record speed. Working with existing customers and successfully onboarding new ones required a steady flow of new talent.
One of the most tactical insights Adam shared was his approach to ramping up new hires. Early on at Scale AI, the company threw people into complex customer relationships without context and paid for it with high turnover and frustrated employees.
Adam helped develop a structured four-week ramp-up plan:
Week 1: Don’t do any work. Just learn about the company and meet people.
Week 2: Start learning about your specific role (think, Account Management, Sales).
Week 3: Learn about your specific account.
Week 4: Begin transitioning to autonomy.
“If you just throw people into the weeds, they typically do not last. And the client suffers, too. You could maybe shorten onboarding to three weeks. But in week one, you should not have people jumping in and starting to prepare client documents.”
Client Delivery as a Services Business
Much has been written recently about client delivery in AI-first companies. A16Z coined it the “The Palantirization of everything”, referring to the Forward-Deployed Engineering model of engineers working closely with customers in project-based settings to provide value and customize existing products. Generally speaking, a core group of engineers and Engagement Leaders or Account Managers would ensure that customers can fully derive value from a product or service.
Working with some of the biggest enterprises in the world, Scale AI followed a similar approach and built up close relationships across their customers. Having managed enterprise relationships at one of AI’s most prominent infrastructure companies, Adam had strong opinions on what separates good Operators from great ones in client-facing roles:
“You are in a services business. Think about going to a nice restaurant: they greet you, they try to make it a welcoming experience. That’s how you need to think about customers. You are representing the company, and you want them to have a fantastic experience.”
Adam also emphasized the importance of being physically present:
“Be on-site with your customers at least once a quarter, ideally twice: once for a mid-quarter visit and once for the QBR. It makes such a difference. I see a lot of younger individuals opting for Zoom calls and my advice has always been the same: hop on a plane and get face time.”
The Future of Human Data
The human data labeling industry, which powers the training of every major AI model, has shifted dramatically since GPT-3.5 went viral in November 2022, Adam remembers:
“Back then, any data you could get your hands on, someone would buy. Now, we’ve moved towards significantly more specialized datasets: advanced math and physics problems, robotics, RL environments, just more complex subjects.”
Adam predicts significant consolidation in the next 12-24 months:
“The barrier to entry historically has been so low. As long as you have a customer, you can find someone to label data. This business has been around for decades. A lot of people saw Alexander Wang be successful and wanted to do the same thing.”
Adam actually sees a much larger opportunity for data labeling in the physical and robotics world. He’s leading with a major question: will companies like Figure and 1X handle data labeling themselves, or go to outside vendors? This and related questions motivated Adam to return to his Engineering roots for his new venture.
Shifting from Operator to Investor
After Scale’s acquisition, Adam decided to formalize his angel investing activities he had been doing for nearly two years. He launched SR² Ventures, and has since invested in 15-20 companies. His investment focus is on startups leveraging human data and robotics while he occasionally helps with hands-on GTM-sparring.
His selection criteria are simple:
“Everyone needs to be a high ARR-per-head business. You need to be a lean organization. If you just throw people at the problem, you won't be able to compete.”
For AI companies specifically, he looks for high-quality, specialized data plays with diversified product offerings:
“Being an AI-wrapper is not enough. You need a unique advantage, for example a competitive moat and paying customers. I see many people building features that do not stand out in the crowd. Maybe you’ll make money for a few months, but it's a matter of time before someone overtakes you.”
Primed by his client-facing work at Scale AI, Adam sees the biggest mistake of early-stage founding teams in building without talking to customers:
“Many companies I speak with base their product decisions on a couple, select conversations. Especially with AI, people get excited about new products easily. But fundamentally, you still need to solve critical problems for the customer. A nice-to-have is very different from solving real business problems, and the willingness to pay long term will differ vastly.”
Leveling up
Asked for his favorite resources and sounding boards, Adam didn’t hesitate and shared the following:
How to Win Friends and Influence People, an all-time classic.
Good to Great for understanding what separates good companies from exceptional ones, a framework he still returns to periodically.
Ben’s Bites, a newsletter for staying up to date in AI.
Thank You
Thanks for reading this issue of The Operator.
I enjoyed learning about Adam’s path from Chief of Staff to Scale AI to running his own angel fund. His framing of client delivery as a “services business” and his structured onboarding approach are frameworks any operator can apply immediately.
Big thanks to Adam Gramling for taking the time to share his journey!
By the way: I’ve just launched my subscriber chat. Would love to see some of you there for discussions and wishes/recommendations for next topics.
See you soon,
Jan