- 🔭 Currently working on Cloud Simulation & AI projects
- 🌱 Learning Cloud Architecture, Deep Learning & DevOps
- 🌍 Contributing upstream to pgmoneta, GRASS GIS, Kubernetes & Flux
- 👯 Open to collaborate on Open Source & Hackathons
- 💬 Ask me about Java, Python, React, Node.js
- 🌐 Check out my portfolio: saipranav.me
- ⚡ Fun fact: I love the rush of hackathons!
Upstream work on C and Python projects — mostly memory-safety and correctness fixes found by running static analysis the projects' CI does not, each with a reproduction and a regression test where the project supports one.
| Project | Contribution | Status |
|---|---|---|
| pgmoneta | Double frees, use-after-free, unchecked allocations, an off-by-one stack overflow, error-path null dereferences, and corner-case tests for the core string helpers | 18 merged · 10 open |
| pgagroal | An unbounded strcat stack overflow in the CLI, unchecked reallocs, and silent truncation in the numeric append helpers — found by writing the corner-case tests |
14 merged · 4 open |
| pgvictoria | Cross-ported allocation checks and string-helper fixes, with corner-case tests for the append family | 7 merged · 3 open |
| pgexporter | Unchecked allocations in the YAML and network paths, plus a kqueue accept-drain fix ported from pgagroal | 10 open |
| grass | Null pointer dereference in the vector library, a null function pointer crash in v.to.rast, 64-bit cell counters, and unbounded environment growth in the runtime setup |
2 merged · 6 open |
| website | Docs fixes, and a style guide section defining deprecated vs no longer served vs removed for APIs | 2 merged · 2 open |
| jabref | Repaired fetcher tests that broke when upstream metadata services changed their responses | 3 merged |
| gnuradio | QA test for the real-time scheduling bindings, and float-tolerance fixes for two numeric tests that compared approximations for bit equality | 2 merged · 1 open |
| hiring-agent | AI agent to evaluate and score resumes. | 2 open |
| flux2 | Removed dead fields from the install flags | 1 open |
Also ml4sci-tasks — ML4SCI evaluation tasks: gravitational lensing substructure classification (ResNet18, 0.9996 accuracy, published alongside the leakage-control check that interrogates it), historical-text OCR with paired bootstrap significance testing, and radio interferometry preprocessing.





