Lan-Anh Dang-Vu

1st Rotation: Data Engineer

Jul. 2025 – Feb. 2026

Wrote and deployed production Python and SQL code to build ETL pipelines using Snowflake, Matillion, and Control-M to automate data ingestion, transformation, and scheduling across systems.
Modernized a data generation and delivery pipeline by re-architecting data workflows and automation, improving reliability and reducing manual processing time.
Wrote SQL-based data validation and test queries for ETL pipelines, systematically checking data quality and catching inconsistencies before they reached downstream reporting.
Debugged and resolved data integrity issues during the production migration of the Asset Cashflows application, ensuring deployment readiness.

2nd Rotation: Analytics Engineer

Feb. 2026 – Aug. 2026

Took technical ownership of a Snowflake-based document ingestion and AI search platform, leading architecture, development, and stakeholder collaboration to support evolving business use cases.
Designed and implemented a RAG pipeline using Snowflake Cortex, Python, and SQL — integrating Snowflake and AWS APIs (including credential/authentication setup and sending/retrieving agent responses) — to process unstructured content and enable scalable AI-powered search across enterprise knowledge bases.
Built automated evaluation and testing frameworks using AWS AgentEval and Snowflake Cortex Agent Evaluations to measure retrieval accuracy, source attribution quality, and response performance for enterprise AI applications.
Developed a CI/CD deployment pipeline in Azure DevOps to automate the release and promotion of Snowflake and application changes across development and testing environments.

3rd Rotation: AI Platform Engineer

Aug. 2026 – Present

Led a team of 7 students on an industry-sponsored project researching and developing AI-generated music detection to protect artists from deepfake copyright infringement.
Developed and trained audio-processing and machine learning pipelines using Python, PyTorch, Google Cloud Platform, librosa, ffmpeg, and mutagen.
Improved RawNet2's performance by integrating a cosine similarity classifier and CLAD loss function, boosting accuracy to 88% on real music and 92% on generated audio, with up to 76% gains on augmented inputs.

Co-authored PoseBench3D, a benchmarking framework that standardizes cross-dataset evaluation of 3D human pose estimation models across the H36M, 3DPW, GPA, and Surreal datasets.
Built object-oriented Python pipelines on Unix-based HPC systems to reproduce and benchmark 18 existing methods, scoring them with the MPJPE and PA-MPJPE metrics.

Utilized SQL in a Postgres environment to construct efficient queries for data profiling, transformation, and analysis. Effectively filtered and organized data sets to derive insights and meet project requirements.
Conducted comprehensive data profiling on over 1,000,000 data points to identify data quality issues and patterns, leading to a significant reduction in data errors.
Created an efficient technique using SQL to reduce download speeds of large data sets from hours to under 5 minutes.
Maintained logical data dictionaries to ensure cohesive documentation.
Collaborated cross-functionally to bridge communication gaps and ensure cohesive work production between teams.

Served on the coding team to program character animations and setting designs using C# and Unity. Attended weekly meetings with the team to brainstorm and pitch ideas to enhance game production.
Collaborated with a team of 5 developers to create 3D environment assets within tight deadlines, meeting all marketing objectives.
Produced the game trailer using Adobe Premiere Pro and Adobe After Effects.

My Résumé