Keith Adams spent nine years as roughly the 20th engineer at VMware before building core infrastructure at Facebook (including founding Facebook AI Research) and serving as Chief Architect at Slack, giving him unusually deep technical experience across multiple iconic platforms. His Google Scholar profile shows highly cited research in virtualization, distributed systems, and AI methods, which is rare depth of original research output for a venture investor.

• Focus on granular systems engineering and backend performance, particularly the mathematics of scalability. He shares detailed insights on how queuing theory impacts server latency, explaining why running servers at lower utilization is often necessary to maintain speed and reliability for interactive applications.
• Enthusiasm for the unglamorous but essential infrastructure that powers AI development. He frequently highlights tools that solve specific friction points for developers, such as reliable PDF parsing for data extraction or solutions that handle the complex dependencies required to run specific AI models.
• Engagement with deep technical theory regarding artificial intelligence and machine learning. He openly muses on complex topics like the relationship between causal calculus and reinforcement learning, or the mechanics of interpreting open-source models like Llama.
• Appreciation for hard tech and rigorous engineering research. He amplifies work that tackles difficult problems, such as distributing Linux kernel bug data or creating manufacturable CAD geometry models, signaling a preference for founders who are deeply technical and detail-oriented.