Summary
DataOps profile with Ph.D in Recommender Systems and skills in Infrastructure and Architecture.
I like the Blockchain concept and the role of technology in entertainment and health. I love sports, photography and contemporary poetry.
Imaginative. Fast-understanding. Polyvalent. Hungry of Knowledge. Open-minded Spanish. French living and English lover. I love to apply what I learnt in the every-day-life.
Experience
eCare targets Smart Homes and collects energy consumption data from smart devices.
Using data, eCare helps Engie Business Units and final users to better understand and control energy consumption making it possible to reduce energy bills.
Skills: AWS, Terraform, Docker, Python, GitHub, Jenkins
The goal is to build a datalake where data integration and transformation is easy and where Data Scientist can easily access data to create models.
Skills: SaltStack, OpenStack, Hortonworks, Kerberos, JupyterHub, Python
Xebia is a consulting firm specialized in IT which advocates for knowledge sharing and quality solutions.
My missions are data-centric but always using a 'dev-ops' approach.
Blackpills is a digital-media startup that produces orignal short-video content and streams them through their own services (mobile apps)
Data, infrastructure and backend architecture team:
1) Defining BigData pipelines
2) Developing services based on collected data to customise user's experience
3) Deployment and maintenance of data and backend services
Skills: ELK, Kafka, Hadoop HDFS, Apache Spark, Cassandra, Couchbase, Docker, Kubernetes, Scala, Bash
Internship in Machine Learning focusing on a very particular issue called 'cold-start': situation in which new users (or items) are new in the system; therefore there is no information about them leading in a miss classification.
Active Learning has been proposed to face cold start by intelligently acquire the needed essential data from very first users.
Skills: Hadoop MapReduce, Java, Machine Learning
Thesis oriented to industry: Thesis dubbed Towards Accurate and Scalable Recommender Systems
The enormous quantity of information in Webs may overwhelm users. Recommender Systems study the interest of users in order to pre-select information in which he/she might be more interested. My goal in this subject is to create a recommender system that relies on two concepts: generic implicit relations in data and distributed collaborative filtering techniques.
Skills: Hadoop MapReduce, Apache Mahout, Java, R-Cran, SQL, Bloom Filters, Weka, Protegee, Jena, SPARQL
Extra: Project Manager for final master projects focused on recommender systems