Victor SONG

[Skills] ► General : Python, SQL, Git, Docker, Terraform, Apache Spark, Apache Airflow, DBT ► GCP : BigQuery, Dataproc, Cloud Workflow, Cloud Storage, Cloud Build, Cloud Run, Looker ► Azure : Databricks, Azure Storage, Azure Functions ► AWS : Redshift, Athena, S3 ► AI/LLM : VertexAI, LangChain, RAG, OpenAI [Projects] ► Built ELT pipelines to handle TB-scale marketing data from Google, Meta, etc. within Cloud Data Warehouse for business teams from 100+ countries. ► Designed and implemented serverless API modules on event-driven data platform for data extraction from divers data providers in parallel. ► Implemented brands matching pipeline using Google LLM in Vertex AI. ► Implemented ETL pipelines to source, transform and persist TB-scale data and deliver E-commerce domain KPIs to Consolidated Data Products. ► Cloud data pipelines migration from Redshift to Spark to improve performance. ► Designed serverless functions on Azure to download, transform and synchronize multilingual documents for GenAI based chatbot and risk assessment application. ► Built LLM-based applications with LangChain, RAG, pgvector and OpenAI API.

Skills : Python, SQL, Git, Neo4j, Elasticsearch, Docker, Terraform, Argo, GCP (DataFlow/Beam, BigQuery, Kubernetes/K8S, Cloud Build, Cloud Storage) ► Developed ETL pipelines to ingest, process, and store TB-scale public and private medical datasets from 100+ data sources to support AI Drug Discovery. ► Implemented AI drug repurposing algorithms and industrialized with web application to provide predictions and insights for pharmaceutical scientists.

Skills : Python, Elasticsearch, Docker, GitLab CI, IBM Cloud, Azure (Databricks/Spark, Data Factory), NLP ► Designed and maintained data pipelines and ML/NLP pipelines for Orange's chatbots for user behavior analysis and product performance monitoring. ► Developed PySpark analytics pipelines and orchestrated with Azure Data Factory. ► Built, evaluated, deployed, and optimized chatbot models in various production environments to meet the needs of 160,000 users in 200+ business scenarios.

Skills : Java, JavaScript, SQL, Oracle DB, SAP Business Object ► Full stack development of Java web applications for visualizing data from 3000+ key accounts globally. ► Performed financial data analysis and automated reporting using SQL and BI to monitor customer performance.

Skills : Elasticsearch, Logstash, Kibana, Python, Shell ► Big Data processing and visualization for Orange's email platform for 2 million users using the ELK stack.