Email | GitHub | LinkedIn | MSR

Pantazis Deligiannis

I am a Principal Research Software Engineer at Microsoft Research,
where I am building new AI-powered tools and techniques for helping developers be more productive and develop more reliable systems.

At Microsoft, I previously created Coyote, an award-winning tool for intelligently testing concurrent C# code and deterministically reproducing bugs. Coyote is used by many core engineering teams across Azure, where it has found tons of critical bugs before they hit production and impact end-users. I have also contributed to several other R&D projects during the past 8 years at Microsoft Research, tackling hard problems across software engineering, programming languages and systems.

I got my PhD in Computing at Imperial College London, where I R&D'ed novel automated-testing techniques. For my thesis, I was fortunate to be advised by Professor Alastair F. Donaldson.

Projects.

I am currently working on:

  • AI4Code: I am building a neuro-symbolic framework that uses AI (LLMs) to automate repository-scale code transformation workflows.

Other cool projects that I have worked on over the years at Microsoft Research:

  • [pdf] RustAssistant, an AI-powered tool that uses large language models (LLMs), such as GPT-3.5 and GPT-4, to automatically fix Rust compilation errors.
  • [git] Coyote, a tool for intelligently testing concurrent C# code and deterministically reproducing bugs. It is used across many core teams in Azure (including Compute, Storage and Networking), as well as Microsoft Teams, and has discovered hundreds of critical bugs before they hit production.
  • [article] Confidential Clean Rooms, a new privacy-preserving distributed programming model, which became part of the Data Empowerment and Protection Architecture (DEPA) standard in India.
  • [pdf] Coyote Actors, a library and runtime for building reliable distributed systems using the actor programming model. It is the production runtime for several core services in Azure Compute. The library interoperates with the Coyote test engine, allowing developers to test their distributed systems, and find and reproduce bugs before deploying to production.
  • [pdf] Reliable State Machines (RSMs), an actor-based runtime for .NET that provides automatic fault-tolerance for distributed systems.
  • [article] Calc.ts, the TypeScript-based calculation engine that powers Microsoft Excel in the browser to accelerate client-side calculations. I designed and implemented a state-machine based architecture for asynchronously evaluating formulas across cells, significantly speeding up calculations.
  • [pdf] Snowflake, an experimental extension of the .NET runtime that provides novel APIs for safe manual memory management alongside the traditional garbage collector.
  • [git] P#, a programming language and runtime (implemented on top of C#) that simplifies designing, developing and testing asynchronous software using asynchronous communicating state-machines. P# started as my internship project at Microsoft Research and then became the main topic of my PhD thesis. P# was used in production as the runtime that powers several core Azure Batch microservices, and has since evolved into Coyote Actors.

Years ago, I had a short stint as an indie game developer:

  • Dungeon Story, an iOS Puzzle/RPG game that I developed for fun in 2012. It has been downloaded by almost 200K users, and has hundreds of 4.5/5 star ratings. It is available worldwide retired from the iOS App Store.

Awards, blog posts, demos and talks.

  • [] Best Software Science Paper Award from the European Association of Software Science and Technology (EASST) for the "Industrial-Strength Controlled Concurrency Testing for C# Programs with Coyote" TACAS'23 paper, presented at the European Joint Conferences on Theory and Practice of Software (ETAPS), April 2023.
  • [video] "Microsoft Research Unplugged – Developer edition", Microsoft Build Conference, May 2021.
  • [video] "Tech Minutes: Project Coyote", Microsoft Innovation Tech Hub, May 2021.
  • [demo] "Explore Project Coyote", Microsoft Innovation Tech Hub, May 2021.
  • [article] "New tools from Microsoft Research help developers write bug-free code, faster", Microsoft AI for Business Blog, May 2021.
  • [] Distinguished Artifact Award from the Association for Computing Machinery (ACM) for the "Learning-based controlled concurrency testing" paper, presented at the International Conference on Object-Oriented Programming, Systems, Languages, and Applications (OOPSLA), November 2020.
  • [article] "Extreme programming meets systematic testing using Coyote", Microsoft Open Source Blog, July 2020.
  • [article] I was honored to be selected as one of the 15 finalists in the UK ICT Pioneers 2015 competition by the UK Engineering and Physical Sciences Research Council (EPSRC) for my PhD work on P#. See this competition video describing my PhD work in three minutes.
  • [] Awarded the Alison Cawsey Memorial Prize for the most deserving MSc student in the 2010-2011 academic session at the Department of Computer Science, Heriot-Watt University, November 2011.

Research.

Peer-reviewed publications, technical reports and theses (dblp | google scholar | orcidORCID iD icon):

  1. "RustAssistant: Using LLMs to Fix Compilation Errors in Rust Code", Pantazis Deligiannis, Akash Lal, Nikita Mehrotra, Rishi Poddar and Aseem Rastogi. In the 47th IEEE/ACM International Conference on Software Engineering (ICSE'25), 2025.
  2. "Leveraging LLMs for Program Verification", Adharsh Kamath, Nausheen Mohammed, Aditya Senthilnathan, Saikat Chakraborty, Pantazis Deligiannis, Shuvendu Lahiri, Akash Lal, Aseem Rastogi, Subhajit Roy and Rahul Sharma. In the 24th International Conference on Formal Methods in Computer-Aided Design (FMCAD'24), 2024.
  3. [pdf] [Best Software Science Paper Award] "Industrial-Strength Controlled Concurrency Testing for C# Programs with Coyote", Pantazis Deligiannis, Aditya Senthilnathan, Fahad Nayyar, Chris Lovett and Akash Lal. In the 29th International Conference on Tools and Algorithms for the Construction and Analysis of Systems (TACAS'23), 2023.
  4. [pdf] "Building Reliable Cloud Services Using Coyote Actors", Pantazis Deligiannis, Narayanan Ganapathy, Akash Lal and Shaz Qadeer. In the 12th ACM Symposium on Cloud Computing (SoCC'21), 2021.
  5. [pdf] "Nekara: Generalized Concurrency Testing", Udit Agarwal, Pantazis Deligiannis, Cheng Huang, Kumseok Jung, Akash Lal, Immad Naseer, Matthew Parkinson, Arun Thangamani, Jyothi Vedurada and Yunpeng Xiao. In the 36th IEEE/ACM International Conference on Automated Software Engineering (ASE'21), 2021.
  6. [pdf] [Distinguished Artifact Award] "Learning-based controlled concurrency testing", Suvam Mukherjee, Pantazis Deligiannis, Arpita Biswas and Akash Lal. In the 35th ACM SIGPLAN International Conference on Object-Oriented Programming, Systems, Languages, and Applications (OOPSLA'20), 2020.
  7. [pdf] "Studying ransomware attacks using web search logs", Chetan Bansal, Pantazis Deligiannis, Chandra Sekhar Maddila and Nikitha Rao. In the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR'20), 2020.
  8. [pdf] "Reliable State Machines: a framework for programming reliable cloud services", Suvam Mukherjee, Nitin John Raj, Krishnan Govindraj, Pantazis Deligiannis, Chandramouleswaran Ravichandran, Akash Lal, Aseem Rastogi and Raja Krishnaswamy. In the 33rd European Conference on Object-Oriented Programming (ECOOP'19), 2019.
  9. [pdf] "Implementing and evaluating candidate-based invariant generation", Adam Betts, Nathan Chong, Pantazis Deligiannis, Alastair F. Donaldson and Jeroen Ketema. In IEEE Transactions on Software Engineering (TSE), 2018.
  10. [pdf] "Project Snowflake: non-blocking safe manual memory management in .NET", Matthew Parkinson, Dimitrios Vytiniotis, Kapil Vaswani, Manuel Costa, Pantazis Deligiannis, Dylan McDermott, Aaron Blankstein and Jonathan Balkind. In the 32th ACM SIGPLAN International Conference on Object-Oriented Programming, Systems, Languages, and Applications (OOPSLA'17), 2017.
  11. [pdf] "Lasso detection using partial state caching", Rashmi Mudduluru, Pantazis Deligiannis, Ankush Desai, Akash Lal and Shaz Qadeer. In the 17th International Conference on Formal Methods in Computer-Aided Design (FMCAD'17), 2017.
  12. [pdf] "Scalable techniques for analysing and testing asynchronous software systems", Pantazis Deligiannis. Ph.D. Thesis, Department of Computing, Imperial College London, 2017.
  13. [pdf] "Uncovering bugs in distributed storage systems during testing (not in production!)", Pantazis Deligiannis, Matt McCutchen, Paul Thomson, Shuo Chen, Alastair F. Donaldson, John Erickson, Cheng Huang, Akash Lal, Rashmi Mudduluru, Shaz Qadeer and Wolfram Schulte. In the 14th USENIX Conference on File and Storage Technologies (FAST'16), 2016.
  14. [pdf] "Fast and precise symbolic analysis of concurrency bugs in device drivers", Pantazis Deligiannis, Alastair F. Donaldson and Zvonimir Rakamaric. In the 30th IEEE/ACM International Conference on Automated Software Engineering (ASE'15), 2015.
  15. [pdf] "Asynchronous programming, analysis and testing with state machines", Pantazis Deligiannis, Alastair F. Donaldson, Jeroen Ketema, Akash Lal and Paul Thomson. In the 36th Annual ACM SIGPLAN Conference on Programming Language Design and Implementation (PLDI'15), 2015.
  16. [pdf] "Automatic verification of data race freedom in device drivers", Pantazis Deligiannis and Alastair F. Donaldson. In the 4th Imperial College Computing Student Workshop (ICCSW'14), 2014.
  17. [website] "Engineering a static verification tool for GPU kernels", Ethel Bardsley, Adam Betts, Nathan Chong, Peter Collingbourne, Pantazis Deligiannis, Alastair F. Donaldson, Jeroen Ketema, Daniel Liew and Shaz Qadeer. In the 26th International Conference on Computer Aided Verification (CAV'14), 2014.
  18. [pdf] "Haskell vs. F# vs. Scala: a high-level language features and parallelism support comparison", Prabhat Totoo, Pantazis Deligiannis and Hans-Wolfgang Loidl. In the 1st ACM SIGPLAN Workshop on Functional High-Performance Computing (FHPC'12), 2012.
  19. [pdf] "Improving the diagnosis of mild hypertrophic cardiomyopathy with MapReduce", Pantazis Deligiannis, Hans-Wolfgang Loidl and Evangelia Kouidi. In the 3rd International Workshop on MapReduce and its Applications (MAPREDUCE'12), 2012.

Professional Activities.

I have served on the Artifact Evaluation committees of the following conferences: ASE 2022, POPL 2017, CGO 2016 and PPoPP 2016.

I have also served as a reviewer for the Transactions on Computer Systems (TOCS) journal.

Other.

I live in WA, USA. I love playing tennis . I also love travelling. In , I have traveled to 15 states: AZ, CA, CO, FL, HI, ID, IL, MT, NE, NM, NV, NY, OR, UT and WA. In , I have traveled to 30 countries: ARE, ARG, AUT, BEL, BHS, CHE, CZE, DEU, ESP, FRA, GBR, GRC, HRV, KOR, LUX, MCO, MNE, MEX, MDV, NLD, IND, ITA, JPN, PRT, SVN, THA, TUR, TWN, USA and VNM. (Country codes in ISO 3166-1 alpha-3.)