Data Science Talk Series 

Language Agents for Software Development (Online)

Abstract: Large Language Models (LLMs) and Generative AI are becoming an integral part of day-to-day tasks at the personal and enterprise level due to their ability to carry out multifaceted language and cognitive tasks. The rapid advancements in LLM research have spurred the development of language agents for a wide array of applications, including writing assistants, translation services, and customer support, all aimed at enhancing human productivity. While language remains a central medium for interaction, the emergence of multimodal inputs and outputs—such as text, code, images, and video—has further broadened the impact of these technologies, reshaping diverse fields.

In particular, the ability of LLMs to generate source code (e.g., Python from natural language) has had a transformative effect on software development, significantly boosting engineering productivity. In this talk, I will discuss the challenges associated with code generation tasks, the methodologies for building effective language agents for program synthesis, their evaluation, and future directions for advancing this field.

Bio of the speaker: Md Rizwan Parvez is a Scientist at the Qatar Computing Research Institute (QCRI), where his research spans natural language processing and generative AI, with a focus on language-code integration, information retrieval, spatial reasoning, and multi-agent modeling. His work has appeared in premier NLP and machine learning venues, including ACL, EMNLP, NAACL, ICML, NeurIPS, and ICLR. His research has received an ICML 2025 Spotlight, been listed in AI-Guides’ most impactful RAG papers, received coverage in outlets such as MarkTechPost and Medium, and accumulated over 3.5 million downloads on Hugging Face. Rizwan earned his Ph.D. in Computer Science from UCLA in 2022, supported by multiple graduate fellowships. Before joining QCRI, he was a Research Scientist at Bosch Research in Sunnyvale, CA, and during his doctoral studies, he completed research internships at Microsoft Research, Salesforce Research, Facebook AI Research (FAIR), and Google Research. Rizwan has served as a Senior Area Chair for ACL 2026, EACL 2026, EMNLP 2025, and  ACL Rolling Review since May 2025.