Remi P.

- Data Platform engineering on Databricks using Terraform

- Developed data pipelines on PySpark and Delta Live Tables (SQL, Python)

- Introduced geospatial capability within the company (Apache Sedona)

- dbt modelling following best practices (SQL models mainly)

- Looker LookML development to build Explores and Views with business stakeholders

- Introduced Python dbt models to enhance page views data

- Built an open-source Python library to sync data warehouse and LookML views

- Built an OCR system to parse unstructured PDF documents

• Data:

o Designed and implemented data quality pipelines for a major telecommunication company in South Africa (Amazon CDK, AWS Glue DataBrew).

o Reduced data processing time from 8 hours to 14 minutes in a Machine Learning leveraging Spark

o Created an Amazon Lex chatbot for an Amazon Connect contact centre for a leading banking group. Increased productivity of agents by 40%.

• MLOps:

o Successfully built a secure multi-account MLOps platform to enable a Data Science team to build new models. Migrated 14 models in three months. (https://github.com/aws-samples/mlops-multi-account-terraform)

o Trained and served XGBoost algorithms on Amazon SageMaker for a large UK insurance group using SageMaker and DevOps best practices (CI/CD using the Code AWS suite). Built a secure custom Docker container for training and serving.

o Created a Terraform template to deploy a secure multi-account MLOps platform on Amazon SageMaker in 10 minutes.

• DevOps:

o Implemented a complex virus scanning solution on AWS using AWS Lambda and S3 for a major Dutch asset management/insurance company