Building software for histopathology, backed by M Ventures and Insight Partners

Led design and implementation of the inference engine's database and user interface, integrated with model selection via MLflow and orchestration through Apache Airflow. Collaborated with medical imaging researchers and implemented features, like streamlining data entry using the Anthropic SDK to parse and post data from medical documents. Automated code linting and pushing Docker images to AWS ECR using GitHub Actions.

https://www.pictorlabs.ai/

Built an AI Agent using Playwright MCP and Claude Code to evaluate AI tissue staining algorithms. Containerized agent using Docker and registered onto AWS ECR to run on NVIDIA Run:ai/Kubernetes clusters. Developed an ETL pipeline with Apache Airflow to retrieve and store data at NFS mounts and S3 buckets and run agent jobs. Engineered a solution to Claude Code’s computer vision inability to tell apart images by providing agentic tools with OpenCV and scikit-image. Integrated authentication and Role-Based Access Control using Auth0 SDK in React and Django Rest Framework, and migrated users from the pre-existing authentication method.

Fall '25 - CS 35L: Software Construction

Balliu Lab

Researching Statistical and Deep Learning methods for gene expression. Reimplemented a multiple hypothesis correction method and statistical tests in C++ using std library algorithms and interfaced using Rcpp library, leading to more than 10x increase in speed over R-based implementation. Developed data pipelines to efficiently run code on 1TB+ real and simulated gene expression data.

Mathematics in Lean by J. Avigad and P. Massot

Read Mathematics in Lean by Jeremy Avigad and Patrick Massot, and learned about dependent type theory. Proved theorems in Algebra and Analysis, like the Schroder-Bernstein Theorem, using Lean tactics. Presented and taught basic concepts in Lean at the Spring DRP Colloquium.