Software Engineer focused on building full-stack products, healthcare systems, AI-assisted workflows and data-driven applications.
ReactNext.jsPythonDjangoFastAPINode.js
Different industries, but the same underlying job: take a process people are holding together manually and turn it into something they can rely on.
Dashboards, admin tools and workflow apps — the screens people use for hours, where a bad layout costs someone their whole afternoon. Usually the work is deciding what to leave out.
Approval flows, operations dashboards, inventory tools
Insurance claims, medical record digitisation and health data that has to line up with real standards. The domain punishes guesswork, which is what makes it interesting.
FHIR · RxNorm · claim settlement workflows
LLM-backed steps inside real workflows — a suggestion a reviewer can accept, automation that removes a manual pass. Advisory by design, never deciding on someone's behalf.
LangChain agents · approval automation
Real-time dashboards and visualisations that turn operational data into a decision someone can make during their shift, instead of a report they read the next morning.
Real-time monitoring · reporting · reconciliation
A few products and systems I've worked on across healthcare, manufacturing, finance, and e-commerce. Each one opens into a full write-up — problem, what I owned, how it worked and what came out of it.
Each figure comes from a specific project rather than a company-wide claim. Hover any card to see where it came from.
30%
Fewer claim processing bottlenecks
360Health — claim settlement workflow
35%
Lower frontend data latency
OEE — API response & state optimisation
25%
Better operational decision-making
Real-time dashboards, React + FastAPI
40%
Increase in sprint velocity
Owning features from design to deployment
50%
Reduction in human error
Digital Permit to Work — paper to digital
3
Developers led
Freelance, October 2022 – August 2023
These are outcomes from individual projects and teams I worked in, not organisation-wide impact. Where a figure is approximate it's written as such in the project write-up.
Three stretches of work, all hands-on. The through-line is owning features from design to deployment and staying close to the people the software is for.
December 2025 – Present
Remote
CurrentInHousen Pvt. Ltd.
Running product development end to end and leading a small team — closer to owning delivery than owning tickets.
August 2023 – June 2025
Hybrid / On-site
Innoart Technologies Pvt. Ltd.
Built React and Python applications across healthcare and chemical manufacturing — claims workflows, healthcare data integrations and real-time operational dashboards.
October 2022 – August 2023
Remote
Freelance
End-to-end web application work for multiple clients — delivery, communication and a small team, not just the code.
B.E. Computer Science and Engineering
Sathyabama University · Chennai, India
2019 – 2023
Not a methodology. Just what I've found holds up across projects.
I want to know what we're trying to achieve, not just what the ticket says. It usually changes the implementation.
I'd rather get a useful version in front of people and iterate than spend a month designing something nobody has touched.
The best technical decision starts with the workflow and the person doing it. Most of my healthcare work was learning the domain before writing code.
Maintainable beats clever. Someone reads this code next — often me, months later, with no memory of why.
Grouped by what I actually use it for. No proficiency bars — happy to talk through any of it in detail instead.
Where most of my product work happens.
Python-first, with Node where it fits better.
Relational by default; document stores when the shape calls for it.
Standards work from claims and medical record digitisation.
Used to remove manual steps, not as a feature for its own sake.
How the work gets shipped.
I'm a Software Engineer with 2+ years of professional experience, and most of what I enjoy sits in the space between product and engineering — working out what a system actually needs to do, then building it.
So far that's meant healthcare insurance claims, real-time factory analytics, permit and approval workflows, POS expense tracking and a B2B commerce platform. Different domains, same shape of problem: a messy real-world process that people are currently holding together with spreadsheets and phone calls.
I move across the stack because the interesting problems don't respect the boundary. A slow dashboard is sometimes a React re-render and sometimes a query; a confusing approval flow is sometimes a UI problem and sometimes a data model that never modelled the real process.
Right now I'm most interested in product engineering, healthcare technology, AI-assisted software and data-heavy applications — the kind of systems where getting the details right is the whole job.
Currently
I'm always interested in interesting products, difficult problems, and teams that care about what they build.
Response time
Usually within a day. Based in India, comfortable across time zones.