jeugregg

Hello,
I am interested in offering my services to your company as a Data Scientist because I think I can develop my skills faster with your huge amount of data and I can help you in this mission and in collaboration with your DATA team. And also, I believe that AI can change the world! So, it is very interesting to be able to democratize Machine Learning in more fields too. I am very motivated to solve your client problems and learn everyday.
Finally, I am convinced that NFTs have a great use that we have not yet imagined.

• Last XP: Web3 data scientist 

After learning blockchain development such as Solidity and explore data on-chain, I have built some Web3 projects study mostly on on-chain data in decentralized finance (DeFi).
I have created a dashboard on Dune Analytics with complex SQL request to monitor Uniswap Liquidity distribution.
I participated in Data Challenges like predict ETH token price or token sentiment for OCEAN Protocol. (Several participations 1st place / 2nd place / 4th place over more than 80 participants) 
To predict ETH, I have developed data pipelines to feed a RNN deep learning model coupled with an auto-encoder model. 

• Deep learning 

Before that I have worked one year for an AI Startup in Seoul, Ailys as a Machine Learning Scientist. 
I have developed in Python an adapted version of RNN models to predict time-series sequences with irregular time steps and missing values. Moreover, I have integrated in the platform some autoML features (e.g. : automatically selecting the most suitable model). 

• PSA Engineer data tools 

Also, I have a long experience of 16 years in Matlab and Python, expert in data tools development for vehicle models & tests bench (PSA Groupe Peugeot Citroen / Stellantis). 
I have created a platform of tools in Matlab/Python for engineer, to facilitate and operate mass simulations (consumption, combustion, mechanics, thermodynamics calculations). I've developed a multi-cpu and multi-host tool to help to dimension the architecture and to optimize the control systems for hybrid and electrical vehicles. 

• Machine learning 

About machine learning models (mainly scikit-learn) in Python, I have done data exploration for e-commerce (Customer Segmentation via Cluster Analysis), extract the most important tags in StackOverflow questions (NLP unsupervised search with LDA, mixed with supervised model). 
Also, I have worked on data exploration for Hospital emergencies to predict load peaks (too many patients at the same time) : study by data visualization with low noise insights, patterns, events frequency. 

• Deep learning 

About deep-learning, I have trained Tensorflow / Keras / pytorch models on different projects: to detect french fake news with fine-tuned BERT model (in NLP field), to predict COVID-19 with Tensorflow LSTM/TCN time-series model, to classify dogs' pictures (with VGG-16, ResNet-50, and an original CNN model). 

• Data engineering 

About Data mining, Data pipeline, and using cloud, I have developed a web app dashboard to follow COVID-19 outbreak in France with Plotly DASH in Python with AWS Lambda (API call to have a prediction) with: a trainedTensorflow Lite model, Docker container on AWS ECR, cluster ECS with EC2 instance, Big Data Spark treatment withEMR (cluster Hadoop on AWS), and finally Airflow to build pipeline to update data for the web app. 

• Optimization 

For a Kaggle competition about Mixed Integer Program Optimization, I have developed a Schedule Optimization withan original Genetic Algorithm method in Python and C++. I have improved algorithm performances with Numba and C++ code in interaction with Python. 

• Background PSA 

When I am involved in a project or a company, I am independent and persevering (I have developed a “home-made” tool to replace a Matlab toolbox for parallel computing which is very expensive). I am very loyal (16 years in the samecompany PSA Groupe Peugeot Citroen carmaker). I can also use my team spirit skill, always trying to answer the user story of an engineer and make the team participate in tools improvement. (I have developed a lot of tools to manage data and models for more than 200 engineers). 

• Tableau/SQL 

I had some experiences with Tableau and SQL to display some insight with on-chain data. 
Also, in my former job position, in France, I had a lot of experience with Excel. But I always tried to use it as a quick data visualization connected with Matlab or as a data source for calculation in Python or Matlab. 

• Full-stack dev app. Android 

I have created a responsive mobile application to manage Axis IP cameras with IONIC v5 AngularJS as Full-stack mobile application developer.

I hope that my profile can fit with your company.