Stay Updated

Stay Updated

Born out of shared research visions at UC Berkeley's RISELab, we continue the lab's tradition of pioneering systems that push the boundaries of data and AI at Disarray.

Our team combines multi-disciplinary research insights and industry experience building production ML systems to create an autonomous ML development solution able to withstand the data disarray of production environments.

Founders

Doris Xin

Co-founder & CEO

Doris has worked in production ML since its early days, using firsthand lessons from deploying real-world models to build systems that tackle the biggest engineering challenges. She was an ML engineer at LinkedIn before the title existed, contributed to Apache Spark MLlib as Databricks' first intern, and earned her PhD at UC Berkeley's RISELab as an NSF Graduate Research Fellow.

Her research revolved around understanding and removing the engineering bottlenecks in ML development and productionization, using large-scale empirical studies of how ML pipelines behave in practice and human-centered research on where automation helps most, so AI can streamline tedious data drudgery while keeping practitioners in control of high-judgment modeling decisions.

At Disarray, she's building the system her research argued should exist: one that addresses fragmented data context, reinvented solutions, and institutional knowledge trapped in people's heads.

Moustafa Abdelbaky

Co-founder & CTO

Moustafa brings two decades of research on autonomous orchestration of complex, unpredictable systems: from distributed systems at Rutgers and IBM Research, where he was a three-time IBM PhD Fellow, to edge-native ML as a postdoc at UC Berkeley RISELab, to serving as a Technical Director at NASA Ames, where he received NASA's Early Career Achievement Medal for secure, real-time AI systems in autonomous aviation and space missions.

Across these domains, he saw the same failure mode: autonomy breaks not because components aren't capable, but because systems can't track what data they used, what code changed, and what environment they're operating in.

At Disarray, he's building autonomous ML engineering systems rooted in the conviction that providing the right context matters more than model capability alone.

Backed by

Our angel investors include DJ Patil (Former U.S. Chief Data Scientist), Anthony Goldbloom (Kaggle founder), and Andy Konwinski (Databricks and Perplexity co-founder).