Md Rizwan Parvez

I am a Scientist at Qatar Computing Research Institute. My research centers on \textbf{language and code} as the foundational interfaces for advancing artificial intelligence (AI). The overarching goal is to enable seamless human–machine interaction by dissolving the boundaries between natural language, computational logic, and dynamic environments. I envision next-generation \textbf{natural language processing (NLP)} systems where language and code jointly serve as a universal medium for communication, reasoning, and coordination among humans, machines, and environments. I am interested in the following five threads

Bridging Natural and Programming Languages: We have developed pioneering works on code generation and competitive problem solving, such as RedCoder, xCodeEval, MapCoder, and CodeSim. Currently, we are focusing more on code reasoning and software engineering (SWE) tasks.

Learning from Open Knowledge and Auxiliary Supervision: We published several impactful papers on retrieval augmented generation (RAG), including FILCO, Open-RAG, that can adaptively determine when to use external retrieval, how to filter noises from retrieval, and leverage them effectively in generation.

Bridging Natural and Programming Languages: We have developed pioneering works on code generation, such as RedCoder, xCodeEval, MapCoder, CodeSim, etc.,

Bridging Natural and Programming Languages: We have developed pioneering works on code generation, such as RedCoder, xCodeEval, MapCoder, CodeSim, etc.,