Sarah Fakhoury
sfakhoury at microsoft dot com GitHub / Google Scholar
Research
I am a Senior Researcher in the Research in Software Engineering (RiSE) group at Microsoft Research in Redmond. In 2022, I received my Ph.D. in Computer Science from Washington State University. My dissertation identified empirical perspectives of developer productivity in practice and introduced a novel framework to model sources of cognitive load during program comprehension in real-working environments.
My research centers around ways to improve the design and evaluation of software development tools to optimize the cognitive effort of programmers. An overarching theme is bridging informal and formal reasoning: taking the intent developers express in natural language, code, and tests, and turning it into specifications machines can check. I am currently interested in leveraging formal methods as a mechanism to provide end-user trust and verifiability of AI-assisted programming tools. My approach is largely interdisciplinary, fusing formal methods, empirical software engineering, machine learning, and human-computer interaction. My work spans several areas:
- AI and Formal Methods
- Using language models to produce specifications, proofs, and provably correct code, and studying how experts write proofs today. [1][2][3][4]
- Specification Inference and Intent Formalization
- Recovering checkable specifications from informal artifacts — issue descriptions, documentation, RFCs, and user interaction. [1][2]
- Testing, Bug Finding, and Repair
- Grounding test generation and defect detection in specification knowledge bases, so that agentic tooling hardens code proactively rather than reactively. I wrote about this work in DeepTest: Systematic Testing Grounded in Specification Knowledge Bases. [1][2][3]
- Developer Cognition and Human Factors
- Measuring what code and tooling actually cost programmers, using eye tracking, neuroimaging, and qualitative study. [1][2][3][4][5][6][7][8][9]
Latest News
| 2026 | Serving as Industry Papers Co-chair for FSE 2026 |
|---|---|
| May 2026 | New blog post: DeepTest: Systematic Testing Grounded in Specification Knowledge Bases — early results from deploying specification-driven testing across dozens of Microsoft components and open-source projects |
| Sep 2025 | New preprint: "Good Vibrations? A Qualitative Study of Co-Creation, Communication, Flow, and Trust in Vibe Coding" |
| Aug 2025 | Check out our latest preprint: What's in a Proof? Analyzing Expert Proof-Writing Processes in F* and Verus |
| Apr 2025 | Two papers accepted to ICSE 2025: "3DGen: AI-Assisted Generation of Provably Correct Binary Format Parsers" and "Towards Neural Synthesis for SMT-Assisted Proof-Oriented Programming" |
Publications
For a complete and up-to-date list of my publications, please check out my Google Scholar profile.