Aarush Agarwal

Search Relevance: Working on machine learning ranking and query-understanding systems for Shopify's commerce search stack, combining semantic retrieval, text-similarity features, and offline/online evaluation to improve product discovery across large, multi-tenant catalogs.

Selected as a Venture Fellow in a highly competitive program focused on leveraging AI and technology for real-world impact.

Conducting startup diligence and market research for the firm, and co-organized VentureHacks, a Felicis x CMU hackathon with $10K+ in prizes, 500+ applicants, and speakers including Felicis partners and a Skild AI founding researcher.

Fraud Detection: Improved buyer-fraud detection accuracy by 3% and reduced training iteration time by 70% through dimensionality reduction, importance-based feature pruning, and BigQuery/Dataflow + Vertex AI pipeline rebuilds.

AI Agent Network: Co-filed a patent for a distributed multi-agent system that decomposes tasks with a Neo4j dependency graph and executes subtasks across specialized agents in parallel.

Sequence Modeling: Designed transformer-based fraud models with embeddings and temporal attention over transaction sequences.

Researching dynamic Mixture-of-Experts architectures under Chenyan Xiong, developing adaptive strategies that expand model capacity on out-of-distribution data while mitigating reasoning degradation in continual pretraining.

Designing autonomous research agents that propose, execute, and evaluate model-adaptation experiments across tasks and modalities including vision, clinical, and financial time-series data, iteratively committing variants that improve reasoning and task-benchmark performance over dense backbones.

First-authored FastGraph, a GPU-resident differentiable k-nearest neighbor algorithm with custom CUDA kernels for low-dimensional graph neural network workflows. FastGraph accelerates graph construction in 2–10D spaces with a bin-partitioned, fully GPU-resident architecture and achieves 20–40× speedups over FAISS, ANNOY, and SCANN.

Engineered PyTorch autograd and gradient operations in C++/CUDA and integrated JIT serialization, reducing KNN runtime by an additional 10% and enabling end-to-end differentiability inside GPU training pipelines.