Software Engineering · M.S. Computer Science at Vanderbilt · Expected April 2027
I build APIs, collaborative applications, storage systems, and developer tools. My work spans Python, TypeScript, and C++, with an interest in reliable systems and machine learning infrastructure.
Seeking 2027 new-graduate software engineering roles. Based in Nashville and open to relocation.
- Two merged PyTorch contributions: C++ NCCL2 all_gather fix #200585 and core validation fix #191092, both with regression tests.
- Deployed software: Proofdesk, used by a professor to publish and update an online textbook.
- Published developer tool: codegraph-engine on npm.
| Project | What I contributed | Evidence |
|---|---|---|
| Proofdesk | Co-developed and maintain a collaborative PreTeXt textbook editor: React/Monaco, Y.js/WebSockets, Node/Express, queued Docker builds, GitHub Actions CI/CD, and a later Pyodide browser preview. | Deployed application · Architecture |
| PulseStream | Built telemetry ingestion and persistence, then addressed batch isolation, offset ordering, and conflicting replays. | 11 API-level failure scenarios; documented local benchmark: 2,933 events/sec persisted during a 30-second send window. Method and limitations |
| CodeGraph | Built a TypeScript/Tree-sitter CLI for dependency graphs, call graphs, and change-impact traversal, with confidence labels for inferred edges. | npm package · Resolution methodology |
| LSM-Tree | Started and maintain a collaborative C++17 storage engine; contribute architecture, design review, and integration, with implementation credit documented for contributors. | 14 CTest programs documented on main; explicit asynchronous vs durable acknowledgment contracts. Durability · Benchmarks |
| Satya | Built a Playwright-based tool that compares UI success claims with API or reload-persisted state; includes a CLI, reports, and FastAPI service. | Seeded-bug demonstrations and validation; production-scale validation remains future work. |
Benchmarks above are local development measurements, not production capacity claims. See each repository for inputs, environment, limitations, and contributor credit.
- Amplify GenAI Innovation Center, Vanderbilt — Research Assistant, Software Engineering · Aug 2026–present
Implemented Projects across React/TypeScript and Python/AWS Lambda: project, file-manifest, and memory APIs backed by three DynamoDB tables, ownership checks, and server-side chat context. Backend PR · Frontend PR. Implementation was tested; release was pending at last confirmation. - VU-BEAM Lab, Vanderbilt — Research Assistant · Oct 2025–Aug 2026
Implemented contrastive pretraining, supervised fine-tuning, and PyTorch data loading and augmentation for ultrasound research. - Vanderbilt Mathematics Department — Research Assistant, Software Engineering · Sep 2025–May 2026
Developed Proofdesk for collaborative textbook authoring. Continued maintaining the project after the appointment, including the browser preview added in June 2026.
- PyTorch distributed C++: fixed NCCL2
all_gathertreating distinct empty tensors as aliased buffers, preventing shape errors during uneven gathers; added distributed regression tests. PR #200585 · Landed commit. - PyTorch core: fixed zero-dimension validation in
torch.unravel_index, preserving empty-input behavior and adding regression tests. PR #191092 · Landed commit. - Research publication: co-author, An Integrated GCN–GAT–AE Framework for Robust Anomaly Detection in Industrial IoT Environments, IEEE CICT 2025.
Languages: C++, Python, TypeScript/JavaScript, Java, SQL
Applications and data: React, Node.js/Express, FastAPI, PostgreSQL/pgvector, DynamoDB, Redis, Kafka/Redpanda, Y.js
Infrastructure and ML: Linux, Docker, Kubernetes, AWS Lambda, GitHub Actions CI/CD, Playwright, PyTorch