About the Candidate
Experienced and results-driven data scientist proficient in NLU-powered chatbots and leveraging open data for business insights. Skilled in Python libraries and Transformer-based models as well as scalable data pre-processing, data-driven decision-making, and designing client-facing dashboards.
Experiencia
Designed and implemented machine learning models and algorithms to solve complex data-driven problems.
Search Engine using Fuzzy Search:
Improved search accuracy by implementing fuzzy matching algorithms, resulting in a 20% increase in search result relevance.
Reduced search query response time by 30% through Joblib multiprocessing optimization and integration with Azure cloud services.
Chatbot for KPMG France:
Developed a production ready Natural Language Understanding (NLU) powered chatbot:
Set performance standards for our transformer models via meticulous training and testing phases.
Established performance benchmarks for transformer-based models through rigorous training, testing, and debugging processes.
Engineered a modular and scalable architecture to seamlessly integrate deep learning models into existing software infrastructure.
Open data project:
Extracted valuable information from open-source data to inform strategic business decisions.
Led a team of 7 engineers in a project focused on leveraging open data to provide companies with improved insights into potential clients.
Extracted data from various sources and utilized Kafka for real-time data streaming.
Implemented three AI modules: spacy NER, topic modeling using Ensemble method for text classification, and spacy matching for optimized text search.
Applied multiprocessing as well as multithreading using concurrence and Joblib to optimize time consumption.
AI consulting for international industry clients.
Designed and optimized a CNN to seamlessly detect traffic signs and monitor bus driver’s emotional and alertness states in real-time with an 80% F1 score.
Joined forces with an automotive industry leader, crafting state-of-the-art solutions to assess drill pin wear using the latest in image analysis techniques.
Created an anomaly detection system for rubber mixer data.
Conducted risk analysis and management utilizing big data analytics.
Developed scalable data pre-processing pipelines to facilitate efficient data analysis.
Provided critical data-driven insights to inform investment decisions.
Oversaw the design of business logic at the core of client-facing dashboards.
Technologies : Python, Pandas, Numpy, Git, Matplotlib
Compartir
Facebook
Twitter
LinkedIn
Telegram
Tumblr
WhatsApp
VK
Mail