Pallavi Bichpuriya | AI Researcher & Data Engineer

About Me

Currently pursuing my Master’s in Computer Science at UNC Charlotte (GPA: 3.83). As a Graduate Research Assistant, I develop text-to-motion pipelines, optimize PyTorch inference, and explore generative AI in 3D motion. Previously mitigated $5M in compliance risk at HSBC by optimizing ETL workflows and migrating data pipelines to AWS and GCP.

Skills

PythonSQLAWSGCPC++Machine LearningMotion ModelingReinforcement LearningREST APIsOpenUSD

Professional Experience

Graduate Research Assistant

AI4Health Center, UNC Charlotte
May 2025 – Present

10x throughput improvement for vision-based experiments via C++17/Python USDZ conversion pipelines.
Deployed text-to-motion AI models on Hugging Face, Google Colab, and Android apps.
Investigating generative AI techniques for multi-modal 3D motion synthesis.

Senior Software Engineer

HSBC, Hyderabad, India
July 2018 – August 2024

Mitigated $5M in compliance risk by resolving 46,500 data gaps across 32 applications.
Boosted campaign metrics by 30% and reduced ETL processing time by 40% through Python and cloud migrations.
Led data validation automation platform, reducing ETL errors by 60%.

Highlighted Projects

USD from glTF

March 2025 – May 2025

Built C++17 library to convert glTF/GLB into USDZ for iOS Quick Look.
Resolved cross-platform CMake build conflicts and managed dependencies.

GitHub →

Masked Motion Model Deployment

Jan 2025 – April 2025

Deployed a text-to-motion AI model across Hugging Face, Android, and Colab.
Trained VQ-VAE and masked-autoregressive Transformer on HumanML3D and KIT-ML datasets.

Demo →

Mental Health Chatbot (RL)

April 2025

Built PPO-based chatbot with mood detection and adaptive reward functions.
Hosted on Hugging Face Spaces using Flask and Gradio.