01 / Profile
I'm a full-stack engineer with a systems habit: I like understanding the data, APIs, UI, and constraints connecting them, not just making one layer look finished. I enjoyed engineering in college in both its forms — learning seriously, and learning with friends through conversations, collaboration, and the occasional late push. I worked hardest when the moment called for it; that balance shaped how I work now.
As I move through the professional world, I'm going back to C to strengthen my fundamentals and spending more time understanding systems at depth, including tools like Git rather than using them at a shallow level. These things take time. That is part of the work, and it is worth it.
02 / Experience
From shipping product features to entering a full-time engineering role, each step has increased my ownership.
Wipro
Jul 2026 — Present
Current role
Joined Wipro as a full-time software engineer following its acquisition of Harman Connected Services. Currently completing the training programme and preparing to contribute at production scale.
Harman Connected Services
Jan 2026 — Jun 2026
Completed
Completed an internship in the Data Engineering team, building a multi-agent system with LangGraph and focusing on the supervisor layer.
Poster.fun
Aug 2025 — Dec 2025
Completed
Built and shipped production features for a design-generation platform, owning the AI image pipeline and authentication infrastructure end-to-end.
03 / Capabilities
A practical engineering practice: understand the system, make the trade-offs visible, and deliver with care.
Build
Think
Use
04 / Work
A small set of systems that show how I think, build, and take problems from idea to implementation.
A practical exercise in concurrency
I built this to understand concurrency in practice. It fetches multiple URLs at once and makes the coordination visible instead of hiding it behind a framework.
What I learned
From UDISE data to a resource plan
The project is a platform for schools to enter resource data from the UDISE portal — number of teachers, washrooms, classrooms, and other facilities. It categorises each school as “odd” or “standard”, then uses an allocation algorithm to identify the resources required to move an odd school toward the standard.
What I learned
Learning retrieval by building the pieces
I built this to understand retrieval instead of treating it as a black box. PDFs are cleaned, split into page-level records, enriched with metadata, embedded, and retrieved for answers.
What I learned
05 / Proof
A few concrete markers of how I learn, contribute, and follow through.
National-level competition · 2024
Built and presented a dynamic resource allocation algorithm for standardizing government schools.
Viewcal.com · 2025 — present
Contributing to a production open-source monorepo used by developers around the world.
ViewLangGraph · MCP · Python
Designing agent coordination, tool access, and human review into useful data workflows.
06 / Writing
Short reflections on systems, learning, and the decisions behind the code.
A practical look at how documents become searchable context, and how cosine similarity helps retrieve a useful answer.
