Thị Mai Phương Phạm

Skills: Commodity Trading

Architected, developed, and managed platforms to improve commodity trading activities throughout the following main projects: - Real time PNL and shaping - Geospatial platform for monitoring different traded commodities given weather conditions - Platform for simulating energy price using parameters linked to energy production, interconnections and weather conditions - Real time power plants’ outage (REMIT) - GME data tool Technical skills included: - Development of microservices and swagger APIs - Python packages management - Design and development of data storage solutions - On-premises - AWS migration - Creation of the messaging application - Creation of ETL pipelines - Infrastructure as a code - Access control - Containerized services with Docker; Kubernetes cluster management - Version control Tools: SQL, GeoJson, MongoDB, Redis, ECS, EKS, ECR, SNS, SQS, Glue, AWS Batch, EventBridge, Git, IAM, S3, Bucket ACL, Signed URL, AAD, Oauth2, Pypi, JFROG, Python, FastAPI, Terraform, Docker, AWS, Azure.

VaR data integration project: Developed a window service application that periodically scans an HBase for updates on the integration chain and notify an ElasticSearch endpoint with the updated data. Developed a Spark job that scans an HBase for trades data integration statuses into an HDFS data lake and periodically compact trades successfully integrated at a prior date. Developed a Spark job that reads data extraction requests from a PostgreSQL db and execute them and log a status into HBase. Migration of feed-api and spark jobs from on-premise to azure. Full-chain backtesting data integration project: Developed a Spring framework based feed api that allows pricers to integrate raw data into an HDFS data lake. The integration status of each entry is updated into an PostgreSQL db. Developed a Spark job that maps raw data of different JSON structures into Parquet files, store them to the HDFS and make them queryable via HIVE. Machine learning project: - Experimented with Markov Chain and LSTM to detect anomalies in risk calculation. Refactored code to allow layered unit testing using DI & composing function. Managed production release, monitored production, provided support for data consumers. Tools: - HDFS, HBase, Hive, SQL, Spark, Kafka, Elasticsearch (Kibana) - Java8, Scala, Maven, Spring boot, Parquet, Swagger, Log4J - Kerberos, LDAP - Testing: Unit test (Mockito), integration test, regression test - XLDeploy, Jenkins, Yarn, Oozie, Git, Bash scripting, Docker - Microsoft Azure

Developed a Scala based API to allow users to subscribe to data feeds: Developed a Web UI for users to login and manage their subscriptions. Implemented an authentication and authorisation layer using LDAP and Ranger. Implemented persistence of subscription details in a PostgreSQL db. Developed a generic Spark job to transfer subscribed data to user’s target directory and scheduled it with Oozie. References/ Project supervisors: - Industrial: L. Gendrier (responsable de pôle Data) - Academic: Professor I. Manolescu-Goujot (École Polytechnique) Tools: - HDFS, HBase, Hive, SQL, Spark - Java8, Scala, orc - Kerberos, LDAP - Yarn, Oozie, Git