E-nno is a Swiss start-up that provides an effective response to the challenges of the energy transition in the real estate sector. As the unique Data Scientist, I'm in charge of the development and the maintenance of end-to-end data science projects dedicated to this energy transition.Main projects:- Development of a heating system optimisation algorithm to reduce the energy consumption of buildings. This algorithm is based on a Tensorflow model and a model predictive control algorithm.- Hydraulic unit failure detection algorithm.- Misconfiguration energy meter detection algorithm.- Development and maintenance of a python library to speed up the development of data science projects and to communicate with the database.- Development of analytics dashboard using Grafana and standalone analysis to help the business team in their day to day decisions.

Within Totsa DigitalLab, in charge of the development, deployment and maintenance of data science services used to automate data flow and to help business operators in their day to day decisions.Main projects:- NER model development to detect deal anomalies.- In charge of the cloud migration of onprem services to AWS.- Web app development to help HR to manage the candidate pool. Use of NLP to rank CV.- Development of an internal framework used to simplify NER models training and serving. Framework composed of multiple python services (Django and FastApi) linked to MLflow, LabelStudio and a Mongo database. UI developed with React.

Development and deployment of machine learning algorithms & expert applications for main industry actors.Main projects:- Oil well pressure prediction used to plan drilling schedule.- Live failure prediction on industrial equipment.- Contribution to an internal Python library development, based on FastApi and Docker, allowing Data Scientist to quickly develop and deploy application to use their predictive models.

Development of a recommendation engine for the NP6 CM platform. Recommendation basedon purchase co-occurence. Algorithm developed with Scala to process the large data volumeavailable.

Development of the XLSTAT software Support Vector Machine module. Improvement of the existing algorithm to allow multiclass classification based on One-versus-One and One-versus-Rest methods. Development done using C++ and VBA.

Publication of a regional energy diagnostic in the industry sector.https://www.slideshare.net/tria-pdl/dreal-pd-l-etude-rgionaleconsonrjindustrie072014