I am Piyush Jha, a third-year Ph.D. student in Computer Science at Georgia Tech, specializing in Artificial Intelligence and Logic under the guidance of Prof. Vijay Ganesh. I work at the intersection of machine learning, logic, and formal methods, with a long-term goal of advancing AI for Science across physics, material science, mathematics, and engineering. My research focuses on building scientifically aligned AI systems that reason, prove, and discover by learning from symbolic structures and certificate-driven feedback.
More recently, I am developing Certification Driven Reinforcement Learning (CDRL), a new paradigm that combines reinforcement learning with formal certificates to accelerate scientific discovery and build reliable learning systems. This work connects symbolic reasoning, planning, and scalable search, with applications in AI for Physics, neutrino model discovery, and material science. Earlier, I developed Reinforcement Learning via Symbolic Feedback (RLSF), where small language models learn complex reasoning by interacting with symbolic tools, and AlphaMapleSAT, a symbolic feedback-driven SAT solver that significantly reduces computation on hard combinatorial problems.
Beyond this, I work on algorithm discovery, combinatorial mathematics, SMT and hardware synthesis, secure and trustworthy AI, and accessibility tools for blind and low vision researchers. My goal is to build AI systems that move beyond pattern recognition toward true reasoning and scientific discovery.
Areas I have worked on before pursuing my Ph.D.:
- Natural Language Processing: Hindi-Sanskrit translation with limited linguistic resources
- Representation Learning: Cross-modal Deep Learning algorithm (CorrMCNN)
- Trustworthy ML: Testing and verification tools for Neural Networks
- AI ∩ Formal Methods: SMT tactic selection
- AI ∩ Software Engineering: Reducing runtime errors in Java programs
- AI ∩ Cybersecurity: LLM in an RL loop to fuzz web applications
- AI ∩ Environmental Science: Predicting seasonal groundwater levels on a nationwide scale
- AI ∩ Law: Legal issues around Deepfakes