Singapore
• Supervised junior Data Scientists and worked cross-functionally with Data Engineering, Data Analytics and IT teams for the design and integration of AI solutions: - Optimizing foreign exchange trading gains with an automated hedging ML strategy. - Predicting the trend of stocks’ prices to generate investment recommendations for customers. - Personalizing the app banners display to customers, increasing conversion rates and boosting revenue. - Delivering personalized app push notifications, improving campaign effectiveness and customer experience. - Generating personalized credit card designs from customer prompts. • Technologies: Python (pandas, scikit-learn, Darts, PyTorch, TensorFlow), LLMs, Spark, Hadoop, Ray, Cloudera
• Worked with stakeholders to identify business needs and designed ML solutions from the ground up. • Developed end-to-end pipelines, including data engineering, training and/or serving for models used to: - Predict the home loan refinancing likelihood for campaign targeting, boosting the revenue. - Classify incoming emails for routing and automatically generate answers, reducing the contact center’s workload. - Compute credit scores from customer features, allowing small loan granting without external credit information. • Designed and developed a set of utilities and frameworks to streamline ML models flows, adopted by teams of data scientists for improving model development efficiency. • Technologies: Python (pandas, scikit-learn, Darts, PyTorch, TensorFlow), LLMs, Spark, Hadoop, Ray, Cloudera
• Developed a natural language question answering system, allowing easily access to information across the company. • Worked on the data engineering and data science aspects of the development of models used to: - Redefine the pricing of the company’s products to increase sales revenue. - Classify and reroute support ticket to improve user experience and resolution time. - Detect valuable business leads from incoming emails to reduce the workload of handling said emails. - Detect anomalies in cloud services usage to prevent unwanted or excessive costs. • Technologies: Python (pandas, scikit-learn, spaCy) LLMs, Knowledge Graphs, Dataiku, Google Cloud Platform
• Developed a model separating the energy consumption of individual appliances from the total energy consumption of a whole house, providing a non-intrusive way to help users measure and understand it. Temporary contract in cooperation with the IRIDIA laboratory at the Université Libre de Bruxelles. • Technologies: Python (pandas, TensorFlow), Deep Learning
• Created an Augmented Reality audio player allowing the user to listen to a music virtually spread in a room. The project was presented as an exhibit at the Sony CLS’s 30th anniversary open house. • Technologies: C, Unity, Vive sensors