BahaEddine Mlaouhi

Conducted workflow analysis and data collection on physician resource allocation in Canadian hospitals to understand real-world optimization constraints.

Designed and developed an AI system to automatically translate real-world resource allocation problems into QUBO formulations.

Implemented QAOA-based quantum optimization using IBM Qiskit, integrating key components such as the transpiler, estimator, sampler, and noise mitigation techniques.

Leveraged D-Wave's quantum annealing hardware to solve the QUBO model and compared performance with gate-based quantum approaches.

Evaluated quantum results against classical optimization benchmarks to assess solution quality and computational advantages.

Designed and developed a tax assistant powered by AI Agents to simplify user onboarding and provide a conversational experience.

Fine-tuned LLMs for human-like, friendly interactions to foster trust and engagement during conversations.

Designed a modular state manager to track user inputs and AI outputs, enabling editable sessions and data collection for model fine-tuning and improvement.

Integrated Retrieval-Augmented Generation to provide accurate and updated responses to general UK tax questions.

Used function calling to delegate complex tax calculations to specialized components, maintaining precision and modularity in system design.

Implemented a reflow mechanism that lets users edit previous answers and auto-updating the conversation and recalculating savings in real time.

Design and development of a URL redirection application to optimize URL management processes.

Exploring the benefits of Spring Boot for efficient backend development.

Development of various features, including URL redirects based on specific rules.

Optimized database performance, improving application speed, reliability, and scalability.

Implemented robust unit testing strategies to ensure high code quality and maintainability.