Nabil Adrar, PhD

𝗔𝗜-𝗗𝗿𝗶𝘃𝗲𝗻 𝗗𝗮𝘁𝗮 𝗠𝗶𝗴𝗿𝗮𝘁𝗶𝗼𝗻 𝗟𝗲𝗮𝗱 (𝗣𝘆𝘁𝗵𝗼𝗻, 𝗝𝗦𝗢𝗡) - 𝗧𝗵𝗮𝗹𝗲𝘀 • Led a PoC assessing feasibility of migrating data from a legacy system to a new platform. • Built an LLM-based framework to automate generation of data conversion logic across heterogeneous JSON structures. • Defined guardrails and validation rules to ensure reliable, consistent LLM outputs in production contexts. • Enabled non-technical users to generate and adapt migration scripts for new data sources efficiently. • Produced a detailed report and cost estimation study to assess feasibility, effort, and scalability of the migration approach. 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁 𝗗𝗲𝘀𝗶𝗴𝗻 𝗮𝗻𝗱 𝗠𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁 (𝗣𝘆𝘁𝗵𝗼𝗻) • Led the full lifecycle development of a GPT-based intelligent agent, from specification to deployment. • Coordinated a multidisciplinary team to integrate AI capabilities with constraint programming solvers. • Designed a system to automatically translate natural language descriptions into mathematical models solvable by LP solvers. • Oversaw testing and validation workflows to ensure accurate and explainable results for end users. 𝗢𝗽𝘁𝗶𝗺𝗶𝘇𝗮𝘁𝗶𝗼𝗻 𝘂𝘀𝗶𝗻𝗴 𝗠𝗮𝗰𝗵𝗶𝗻𝗲 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗮𝗻𝗱 𝗖𝗦𝗣 𝗧𝗲𝗰𝗵𝗻𝗶𝗾𝘂𝗲𝘀 (𝗖++) • Developed a ML-driven module for optimal tree decomposition, enhancing graph preprocessing. • Trained predictive models using historical decomposition data to guide CSP solving strategies. • Improved computational efficiency and scalability of CSP resolution via hybrid approaches.