About Me
I am a research scientist passionate about advancing AI's reasoning, problem-solving, and general intelligence, with a strong emphasis on interpretability and alignment. My work integrates formal methods, mathematical logic, and program synthesis to design AI systems that are not only interpretable but also capable of reasoning effectively, generalizing across diverse domains, and addressing complex, open-ended problems.
My primary interest lies in developing AI systems that emulate human reasoning, bridging the divide between abstract theoretical frameworks and practical advancements in general intelligence—all while ensuring alignment with human values. I am always open to collaboration and meaningful discussions about the future of intelligent systems.
Employment
-
Research Scientist: Software Engineering Institute | Carnegie Mellon University
May 2023 to Current. -
Postdoctoral Researcher: Tufts University , under Prof. Jeff Foster
Synthesizing programs from modular program blocks, using row-type inferencing, symbolic solving, and genetic programming.
August 2020 to September 2021. -
Research Internship: Runtime Verification
Verification of Solidity bytecode for smart contracts used in the Ethereum Blockchain.
June 2020 to August 2020.
Education
-
Integrated Ph.D., Computer Science
University of Illinois at Urbana-Champaign, December 2019
-
Master of Science, Computer Science
Chennai Mathematical Institute, India, August 2011
GPA: 3.62 / 4
Publications
-
Synthesizing Contracts Correct Modulo a Test Generator
PDF
with Angello Astorga, Ahmad Dinkins, Felica Wang, P. Madhusudan, Tao Xie
ACM SIGPLAN International Conference on Object-Oriented Programming, Systems, Languages, and Applications (OOPSLA), 2021. -
A Learning-Based Approach to Synthesizing Invariants for Incomplete Verification Engines
PDF
with Daniel Neider, P. Madhusudan, Pranav Garg, Daejun Park
Journal of Automated Reasoning (JAR), 2020. Invited Article -
Learning Stateful Preconditions Modulo a Test Generator
PDF
with Angello Astorga, P. Madhusudan, Shiyu Wang, Tao Xie
ACM SIGPLAN Conference on Programming Language Design and Implementation (PLDI), 2019. -
Sorcar: Property-Driven Algorithms for Learning Conjunctive Invariants
PDF
with Daniel Neider, Pranav Garg, P. Madhusudan
International Static Analysis Symposium (SAS), 2019. -
Compositional Synthesis of Piece-Wise Functions by Learning Classifiers
PDF
with Daniel Neider, P. Madhusudan
ACM Transactions on Computational Logic (TOCL), 2018. Invited Article -
A Decidable Fragment of Second Order Logic With Applications to Synthesis
PDF
with P. Madhusudan, Umang Mathur, Mahesh Viswanathan
EACSL Conference Computer Science Logic (CSL), 2018. -
Invariant Synthesis for Incomplete Verification Engines
PDF
with Daniel Neider, Pranav Garg, P. Madhusudan, Daejun Park
International Conference on Tools and Algorithms for the Construction and Analysis of Systems (TACAS), 2018. -
Synthesizing Piece-Wise Functions by Learning Classifiers
PDF
with Daniel Neider, P. Madhusudan
International Conference on Tools and Algorithms for the Construction and Analysis of Systems (TACAS), 2016. -
Alchemist: Learning Guarded Affine Functions
PDF
with Pranav Garg, P. Madhusudan
International Conference on Computer Aided Verification (CAV), 2015. -
NetGen: Synthesizing Data-plane Configurations for Network Policies
PDF
with Santhosh Prabhu, P. Madhusudan
ACM SIGCOMM Symposium on Software Defined Networking Research (SOSR), 2015. -
Syntax-Guided Synthesis
PDF
with Rajeev Alur, Rastislav Bodik, Eric Dallal, Dana Fisman, Pranav Garg, Garvit Juniwal, Hadas KressGazit, P. Madhusudan, Milo M. K. Martin, Mukund Raghothaman, Sanjit A. Seshia, Rishabh Singh, Armando Solar-Lezama, Emina Torlak, Abhishek Udupa
Dependable Software Systems Engineering, NATO Science for Peace, Security Series., 2015. Invited Article
Thesis
-
Learning Frameworks for Program Synthesis
PDF
Ph.D. Dissertation, University of Illinois at Urbana-Champaign, 2019
Advisor: Prof. Madhusudan Parthasarathy -
A Survey of Automata and Logics Over Infinite Graphs
PDF
Master's Thesis, Chennai Mathematical Institute, 2011
Advisor: Prof. K. Narayan Kumar
Tools
-
Precis
: Synthesizes postconditions of C# programs using the test generator PEX. -
Provisio
: Synthesizes preconditions of C# programs using a test generator as the verification oracle. -
Sorcar
: Algorithm to learn conjunctive inductive invariants. -
NPI
: Synthesizes inductive invariants using an incomplete verification oracle. -
Alchemist
: A General-purpose SyGuS solver to synthesize CLIA expressions.
Won second place in the 2015 SyGuS competition (CLIA track). -
NetGen
: Synthesizes network data-plane configuration changes from high-level policy change specifications.
Teaching
-
Formal Software Development Methods (CS477):
Fall 2016
University of Illinois at Urbana Champaign. -
Software Design Studio (CS126):
Fall 2017, Spring 2018
University of Illinois at Urbana Champaign. -
Data Visualization (CS498 DDV):
Summer 2018
Part of Data Mining Specialization at Coursera, offered by University of Illinois.
Talks
-
Excape PI Meeting, 2017:
Network Synthesis.
-
Midwest Verification Week, 2015:
Network Synthesis.
-
CAV, 2015:
Alchemist: Learning Guarded Affine Functions.
-
SOSR, 2015:
NetGen: Synthesizing Data-plane Configurations for Network Policies.
-
Excape PI Meeting, 2015:
NetGen: Synthesizing Data-plane Configurations for Network Policies.