From Research Lab to Record-Breaking Product: How OpenAI Engineered for Unprecedented Scale
with Sulman Choudhry, Samir Ahmed & Lawrence Bruhmeller
This is a special episode, highlighting a session from ELC Annual 2025! OpenAI evolved from a pure research lab into the fastest-growing product in history, scaling from 100 million to 700 million weekly users in record time. In this episode, we deconstruct the organizational design choices and cultural bets that enabled this unprecedented velocity. We explore what it means to hire "extreme generalists," how AI-native interns are redefining productivity, and the real-time trade-offs made during the world's largest product launches. Featuring Sulman Choudhry (Head of ChatGPT Engineering) and Samir Ahmed (Technical Lead), moderated by Lawrence Bruhmeller (Eng Management @ Sigma).
ABOUT SULMAN CHOUDHRY
Sulman leads ChatGPT Engineering at OpenAI, driving the development and scaling of one of the world’s most impactful AI products. He pushes the boundaries of innovation by turning cutting‑edge research into practical, accessible tools that transform how people interact with technology. Previously at Meta, Sulman founded and scaled Instagram Reels, IGTV, and Instagram Labs, and helped lead the early development of Instagram Stories.
He also brought MetaAI to Instagram and Messenger, integrating generative AI into experiences used by billions. Earlier in his career, Sulman was on the founding team that built and launched UberEATS from the ground up, helping turn it into a global food delivery platform. With a track record of marrying technical vision, product strategy, and large‑scale execution, Sulman focuses on building products that meaningfully change how people live, work, and connect.
ABOUT SAMIR AHMED
Samir is the Technical Lead for ChatGPT at OpenAI, where he currently leads the Personalization and Memory efforts to scale adaptive, useful, and human-centered product experiences to over 700 million users. He works broadly across the OpenAI stack—including mobile, web, services, systems, inference, and product research infrastructure.
Previously, Samir spent nine years at Snap, working across Ads, AR, Content, and Growth. He led some of the company’s most critical technical initiatives, including founding and scaling the machine learning platform that powered nearly all Ads, Content, and AR workloads, handling tens of billions of requests and trillions of inferences daily.
ABOUT LAWRENCE BRUHMELLER
Lawrence Bruhmuller has over 20 years of experience in engineering management, much of it as an overall head of engineering. Previous roles include CTO/VPE roles at Great Expectations, Pave, Optimizely, and WeWork. He is currently leading the core query compiler and serving teams at Sigma Computing, the industry leading business analytics company.
Lawrence is passionate about the intersection of engineering management and the growth stage of startups. He has written extensively on engineering leadership (https://lbruhmuller.medium.com/), including how to best evolve and mature engineering organizations before, during and after these growth phases. He enjoys advising and mentoring other engineering leaders in his spare time.
Lawrence holds a Bachelors and Masters in Mathematics and Engineering from Harvey Mudd College. He lives in Oakland, California, with his wife and their three daughters.
This episode is brought to you by Span!
Span is the AI-native developer intelligence platform bringing clarity to engineering organizations with a holistic, human-centered approach to developer productivity.
If you want a complete picture of your engineering impact and health, drive high performance, and make smarter business decisions…
Go to Span.app to learn more!
SHOW NOTES:
- From research lab to record-breaking product: Navigating the fastest growth in history (4:03)
- Unpredictable scaling: Handling growth spurts of one million users every hour (5:20)
- Cross-stack collaboration: How Android, systems, and GPU engineers solve crises together (7:06)
- The magic of trade-offs: Aligning the team on outcomes like service uptime vs. broad availability (7:57)
- Why throwing models "over the wall" failed and how OpenAI structures virtual teams (11:17)
- Lessons from OpenAI’s first intern class: Why AI-native new grads are crushing expectations (13:41)
- Non-hierarchical culture: Using the "Member of Technical Staff" title to blur the lines of expertise (15:37)
- AI-native engineering: When massive code generation starts breaking traditional CI/CD systems (16:21)
- Asynchronous workflows: Using coding agents to reduce two-hour investigations to 15 minutes (17:35)
- The mindset shift: How rapid model improvements changed how leaders audit and trust code (19:00)
- Predicting success: "Vibes-based" decision making and iterative low-key research previews (20:43)
- Hiring for high variance: Why unconventional backgrounds lead to high-potential engineering hires (22:09)
LINKS AND RESOURCES
This episode wouldn’t have been possible without the help of our incredible production team:
Patrick Gallagher - Producer & Co-Host
Jerry Li - Co-Host
Noah Olberding - Associate Producer, Audio & Video Editor https://www.linkedin.com/in/noah-olberding/
Dan Overheim - Audio Engineer, Dan’s also an avid 3D printer - https://www.bnd3d.com/
Ellie Coggins Angus - Copywriter, Check out her other work at https://elliecoggins.com/about/
