Pranav Rajaram

Data Science, UC San Diego

Experience

ShopifyData Science InternSep 2026 – Dec 2026

Building data models and analytics tools to support product insights on the Checkout & Accounts team.

KlaviyoSoftware Engineer InternJun 2026 – Aug 2026

Built an async JSON:API REST service in FastAPI with full CRUD endpoints, gRPC backend integration, cursor pagination, Kubernetes deployment, and Sentry alerting. Reduced API response latency by 20% using Redis caching.
Developed a full-stack framework with React, TypeScript, and Python that enabled customers to create, configure, and deploy custom object sorting algorithms, decreasing configuration time by 35%.
Implemented a collaborative filtering recommender system using implicit to personalize custom object feed results based on user engagement signals, improving customer feed relevance by 5x.

Oklahoma City ThunderData Science InternJan 2026 – May 2026

Built a GRU-based sequence model with attention in TensorFlow on NBA pose tracking data to predict bad-pass steal probability, improving postgame analysis quality by 15%. Stored insights in a dynamic R Shiny dashboard.
Engineered spatiotemporal features from raw player tracking data and addressed class imbalance with weighted loss, achieving a 25× improvement in PR-AUC and log loss over baseline.
Deployed the model end-to-end on AWS SageMaker and developed scalable S3 pipelines to enable real-time inference on live in-game data.

VeriskData Science InternJun 2025 – Aug 2025

Built a multi-agent RAG workflow with LangChain to simulate legal processes for casualty claims, reducing research time by 40% and improving consistency in settlement evaluation.
Developed an NLP evaluation pipeline (keyword overlap, semantic similarity) to evaluate agent responses, improving accuracy of legal document alignment scoring.
Implemented Monte Carlo simulations to model distributions of claim settlement outcomes, enhancing risk evaluation and leading to more precise premium-setting strategies.

ShotQualityData Science InternJul 2024 – Dec 2024

Utilized Python and SQL to perform rigorous quality checks on basketball tracking data, leading to a 20% improvement in data accuracy for shot probability models.
Enhanced on-court player detection with bounding box refinement using PyTorch, improving tracking accuracy by 15% and enabling new player movement insights.
Created and deployed a dynamic R Shiny dashboard connected to AWS S3 for real-time data quality monitoring.

Yahoo & 4for4Data Analyst & ColumnistJun 2021 – Aug 2025

Authored 100+ data-driven articles read by 5,000+ monthly subscribers at 4for4, over 100,000 readers at Yahoo.
Developed NFL Lookahead Lines web application using R Shiny, providing interactive betting line insights.