Augmenta
A privacy layer for LLM workflows with PII detection, anonymization, and service-oriented integration — built around privacy tooling, API design, and infrastructure-oriented applied AI.
PrivacyLLMPythonFastAPIDockerApplied AI
Selected Projects
A focused set of technical projects across ML, applied AI, and systems engineering.
Energy Demand Forecasting
A 24-hour ahead electricity demand forecasting pipeline with feature engineering, rolling-origin validation, tree-based models, and uncertainty estimation — focused on reproducible ML evaluation.
Machine LearningTime SeriesPythonScikit-learnForecastingXGBoost
Hybrid Classical–Quantum Neural Networks for Audio Emotion Classification
An end-to-end speech emotion recognition pipeline on CREMA-D using mel-spectrograms, transfer learning, and hybrid quantum/classical heads — focused on controlled experimentation under realistic resource constraints.
Quantum MLPyTorchPennyLaneSpeech ProcessingDeep LearningCREMA-D
Support Ticket Classifier
An AI-assisted support ticket classifier with validated structured outputs, deterministic fallbacks, and a lightweight local UI — focused on maintainable backend design and practical handling of model outputs.
AINLPPythonPydanticAnthropicClassification