Pranav Belligundu

B.S. Statistics & Data Science & B.S. Mathematics · Minor in Computer Science

The University of Texas at Austin · Austin, TX

I'm an undergraduate researcher at UT Austin pursuing a double degree in Statistics & Data Science and Mathematics with a Minor in Computer Science. My research, conducted at the IDEAL Lab under Prof. Joydeep Ghosh, focuses on goal-conditioned reinforcement learning for personalized recommender systems. More broadly, I am interested in contributing to frontier ML/AI — spanning physical intelligence, reinforcement learning, and large language models. I've worked across academic research and industry in ML engineering and software development. Outside of work, I enjoy sports (volleyball and soccer), thrifting, and cooking. I also love films (Letterboxd) and shoot film photography (@belli.cam).

Research Interests

Physical Intelligence Reinforcement Learning Frontier LLMs Recommender Systems Retrieval-Augmented Generation Agentic AI AGI Machine Learning Natural Language Processing

Education

The University of Texas at Austin

B.S. Statistics & Data Science & B.S. Mathematics (Double Degree)
Minor in Computer Science

Aug 2023 – May 2027 (expected)

Monta Vista High School

High School Diploma · Cupertino, CA

Experience

Machine Learning Engineer Intern — Workday

Incoming Summer 2026 · Agentic AI · AI Accelerate Team

Machine Learning Undergraduate Researcher — UT Austin IDEAL Lab

Sept 2025 – Present · Austin, TX

Research under Prof. Joydeep Ghosh and Shijun Li on goal-conditioned RL for personalized recommender systems
Model trained with PPO and retrieved-token masking to autonomously query user histories and item metadata
Ran large-scale CUDA experiments while fine-tuning parameters to improve accuracy and overall training quality

Software Engineering Intern — Briq

Jun 2025 – Aug 2025 · Remote

Developed and deployed integration tests across 10+ microservices for an AI-powered invoice processing platform
Automated REST API validation with Postman, verifying 100+ document payloads, cutting QA time by 60%
Reduced release rollbacks 20% through GitLab CI/CD pipeline optimizations that improved deployment reliability

Machine Learning Engineer Intern — Innoflexion

May 2025 – Jul 2025 · San Francisco, CA

Led two interns to develop a feature in DeepRoot, a platform that prepares enterprise data for GenAI use cases
Developed an MoE architecture that routed structured and unstructured data across 30+ LLMs/SLMs with vLLM
Formulated Data Readiness Index framework that evaluated quality, structure, security, and fairness of input data

Data Science Research Intern / TA — Stanford Medicine

May 2022 – Jan 2024 · Palo Alto, CA

Processed and analyzed RNA-seq data from 76 osteoarthritis patients in R to determine valuable biological insights
Used t-SNE, SVA, and UMAP to extract biological signals and identified 13 genes associated with osteoarthritis
Designed an optimized single-cell RNA-seq workflow and wrote a paper documenting the methodology/results

Papers

Integrating Collaborative and Meta-Information Retrieval with Reasoning: An End-to-End Framework for LLM-Based Recommendation

Shijun Li, Joydeep Ghosh, Pranav Belligundu

In Progress

Selected Projects

Apanto

LLM router that intelligently routes prompts across commercial APIs and Hugging Face models. Full-stack app (React + FastAPI) tracking latency, cost, and usage across 20+ models.

Python · React · FastAPI · PostgreSQL · MongoDB · vLLM · CUDA

View on GitHub

VitaRize · 2nd / 160 at Cisco BridgeHacks

Accessibility app that summarizes articles with text-to-speech for visually impaired individuals. Marketed to 10+ users improving content accessibility.

Python · Streamlit · PyTorch · Googletrans

View on GitHub

Medimatch

ML model to predict and detect sicknesses based on patient markers using Decision Trees and Random Forest.

Python · Scikit-learn · Pandas

View on GitHub