azadeh kakavand

Developed and deployed an LLM pipeline using LangChain and GPT-4o to process large sets of unstructured PDF documents. Applied NLP techniques to extract, compare, and recommend information. Integrated of LLMs into applications. Implemented a Speech-to-Text pipeline to automatically fill forms from voice input. Tools: Python, Pandas, Numpy, Seaborn, LangChain, LLMs, NLP, spaCy, NLTK, OpenAI GPT4o, Whisper, Docker, Git, FastAPI, Flask

- Teaching Mathematics and Python programming cours. - Université Léonard de Vinci, La défense - Paris, Computer Vision & Image Processing - Project – Computer Vision (Object Detection and Image Segmentation) Tools: Python, Deep Learning libraries (PyTorch)

Improvement of the user experience for the 'Runtastic Pedometer' application. Integration of the Sepia principle into the health application. Implementation of marketing strategies in the app market.

Development of a Brain-Computer Interface (BCI). Improvement of the accuracy of implemented models. Conducting numerical simulations and calculations. Implementation of neural networks (RNN/CNN/LSTM). Developed machine learning models for decoding EEG signals in a Brain-Computer Interface (BCI). Development and optimization of end-to-end ML pipelines. Performed simulations and numerical experiments. Improvement of the accuracy of implemented models. Tools: Python, Pandas, Numpy, Matplotlib, Keras, Scikit-Learn

Improvement of algorithmic performance (MATLAB application). Implementation of Riemannian geometry in the developed application. Tools: Matlab

Optimization of processes/calculations to maximize machine performance. Tools: Python