Text-to-Speech System with Real-time Voice Cloning
An API for producing high-quality speech from written text with the ability to select different voices and clone voices from three seconds of audio. The model was developed as a result of my master's research. I completed all development stages; data collection, data processing, model development, packaging, and deployment.
Customer Oriented Chatbot
A chatbot developed with the MS Bot Framework that answers the client's core questions and re-routes to the appropriate support agent based on the intent. I proposed the product after doing a comprehensive company business analysis: this improved conversion rates and customer support quality. I designed and implemented software architecture, trained needed NLP models, and integrated the application into the company software solution.
Notification System for Anomalies in Blood Pressure
An API that predicts anomalies in blood pressure and allows medical workers to take preventive measures. The solution was developed for the medical institution of an Arab country. After three months of research and rapid prototyping, I delivered a model that satisfied the requirements for further funding. I also developed a lightweight API for the MVP version.
Analysis Tools of Feedback from Social Platforms
This solution collects posts from social media channels regarding the company, analyzes them, then visualizes them for BI specialists. I developed a complex automated data pipeline. For analysis, I created NLP models using TensorFlow and visualized key insights in an R Shiny-based dashboard.
Marketing Attribution Strategy
Defined custom marketing attribution strategy, which incorporated click-based, incrementality-based, and ML-based attribution approaches. I designed software architecture, implemented the model with SQL and Python, and actively took part in research and presentation stages.
Chatbot For Automation Product
A chatbot developed with the MS Bot Framework in Python that translates client messages into specific actions and answers product-related questions. The chatbot was able to translate user messages into specific SQL queries. I was responsible for the project architecture and implementation; cognitive and AI capabilities of the solution; deployment and CI/CD setup.
Map Of Historical Artifacts Migration
An open-source project that collected data from Wikipedia and used PySpark in combination with various NLP models to build a map of historical artifacts and their movement.I was responsible for project architecture, data collection and cleaning, and Deep Learning models used for entity extraction and linking.
