Computer Science Undergraduate @ Scaler School of Technology | BS in Data Science & Applications @ IIT Madras
I am focused on software engineering, AI/ML systems, backend development, and open-source scientific Python.
I learn by building projects, investigating real-world bugs, contributing to existing codebases, and documenting the engineering decisions behind my work.
I document the problem, investigation, implementation, and lessons learned from each contribution in my Work Journal.
-
SciPy · PR #26540 · Merged
Added shape validation for the inverse covariance matrixVIinscipy.spatial.distance.pdistandcdistwhen using the Mahalanobis metric. The validation prevents incompatible matrix shapes from reaching the lower-level distance kernel, where they could cause buffer over-reads or silent arithmetic errors. Added regression tests for both functions. -
statsmodels · PR #10704 · In Review
Fixed handling of zero observed counts inpowerdiscrepancy$G$ -test calculations, usingscipy.special.xlogyto handle mathematical limits cleanly and avoidNaNstatistics and divide-by-zero warnings.
| Project | Technologies | What it explores |
|---|---|---|
| ProofGraph | Python, NetworkX, LLMs | Evidence-backed claim verification: atomic claim extraction, source retrieval, evidence assessment, and graph-based relationships between claims and sources. |
| IncidentLens | Python, FastAPI, Hindsight, Groq | An incident-investigation assistant that uses persistent memory to retrieve relevant historical incidents and support troubleshooting. |
| Pulse Incident Engine | React, Node.js, Socket.IO, TanStack Query, MongoDB | Real-time incident management application with explicit lifecycle transitions, hash-based deduplication, and live updates. |
| ContextMesh | Python, ChromaDB, Groq, Whisper | A multimodal RAG system for retrieving information from PDFs, images, audio, and video, with source and timestamp grounding. |
- Languages: Python, Java, JavaScript, SQL
- Backend & Application Development: FastAPI, Node.js, Express, REST APIs, WebSockets
- Data & AI/ML: NumPy, pandas, SciPy, scikit-learn, PyTorch, ChromaDB, NetworkX
- Databases & Tools: PostgreSQL, MongoDB, Git, Linux, Docker, Postman
- Work on real codebases and investigate bugs instead of relying only on tutorials.
- Practice implementing mathematical and algorithmic concepts from scratch.
- Participate in open-source contributions and learn through community code reviews.
- Document technical decisions, debugging processes, and lessons in my Work Journal.