My daily work consists of understanding customers' main problems, brainstorming with my team to know how data science could solve them, implementing the solution by writing production-ready code, and presenting the outcomes compellingly.

My main achievements include:

• Implemented an NLP algorithm to extract the sentences and charts mostly linked to a given topic from a scientific paper. Computed the similarity between every sentence and the topic through a word2vec model and highlighted only the most similar ones. Empowered a European agency to reduce the time needed for the literature review by 80%.

• Enabled an international bank to estimate the total income of customers with only secondary accounts. Extracted 600 features from personal and transactional data through Spark and reduced dimensionality with wrapper techniques. Managed data sparsity by creating 9 clusters of users and by making 9 distinct XGBoost models with each one having its own populated features. Reduced the MAPE by 21% compared to the previous estimate.

• Built a random forest model to predict sales of the following 6 months of all products for a multinational fashion company to ensure not to produce too many items. Managed predictions of new products without a history by comparing them with old products with a K-NN model. Coded a POC in AWS, obtained a MAPE of 15% on the first 10 tested items and decided to industrialize the model globally.

• Currently leading a team of 5 data scientists/engineers. Actively mentoring them to express their potential and deliver high-quality results. Always worked with other technical teams, i.e. BI and strategy, and clients from 5 different industries. Conducted over 50 presentations with the usage of video communication tools like Prezi to make them clear and compelling. Directly participating in defining marketing strategies and thinking about making propositions stand out from competitors' ones.

As a Pwc team member, I daily keept me updated about the most innovative technologies and scientific papers and cooperated with my team to implement them to impact business significantly.

I daily cooperated with my team to implement the most innovative technologies to impact our customers business significantly.

My main achievements include:

• Optimized budgeting for a multi-utility company by decreasing the monthly gas demand prediction MAPE by 12% by implementing a random forest model. Managed features generation from a 2-billions-rows distributed DB through Spark. Extended the dataset with meteorological data by building a web-scraping script to download them monthly. Provided stakeholders with 5 reports in the cloud to understand the contribution of every external and internal factor in the predictions through features importance analysis.

• Built a logistic regression model to compute the 3-months churn probability of all employees of a fashion company. Leveraged the trained model’s parameters to identify employees' sources of discontent and give the HR team suggestions to fix the issues in advance. Reduced the turnover rate by 8% in 6 months.

• Enhanced the email marketing campaign of a primary fashion company. Designed an innovative neural network that made them find the next best offer for every customer, starting from their personal data and purchase history by using state-of-the-art text generation models. Increased the campaign conversion rate by 8% in 6 months.

• Reduced by 95% the time needed for a waste management company to produce documentation. Built from scratch a data warehouse and an ETL process to collect in a graph DB and visualize the whole waste path from producer to the last recipient.

• Designed and delivered an interactive dashboard in Sap Analytics Cloud to allow a small appliance manufacturer company to analyze sales data in real-time. Merged sales data with cost data to compute and visualize profit and loss with a city and object detail. Worked with stakeholders to build trust in the tool and to create a data drive culture.