/STORY — FIELD NOTES
A field log of the work, research, experiments, and small decisions that shaped how I build.
CH 01 · Origins
2015 – 17
FN-HIN · 2015 – 17
Prathamic to Praveen, eight stages
Artifact: Dakshina Bharat Hindi Prachar Sabha examinations
Chapter 01 · 2015 – 2018
School in Salem — language study, environmental work, and early community leadership.
Artifact: Dakshina Bharat Hindi Prachar Sabha examinations
Started with Prathamic in 2015 and completed the full eight-examination progression through Praveen Uttarardh, earning First Class at every level.
Lesson — Long progress is easier to trust when every stage has a finish line.
Artifact: Vice President, National Green Corps
At Sri Vidya Mandir, our National Green Corps team grew an organic vegetable garden, donated more than 1,000 saplings to the Isha Foundation, and organised a district-level environmental awareness event. Seeing the first harvest change scepticism into participation made the work unforgettable.
Lesson — Visible progress turns a good cause into shared ownership.
Artifact: Club director role
Served as a director in Rotaract during school. It was an early lesson in showing up for shared work, organising with peers, and treating community service as a practice rather than a badge.
Lesson — Responsibility becomes real when other people are counting on you.
Chapter 02 · 2018 – 2022
Computer Science at Kongu — coursework, research, peer mentorship, and student coordination.
Artifact: Admission letter
Started B.E. Computer Science at Kongu Engineering College. Coursework built the fundamentals; side projects made them practical.
Lesson — The coursework and the side projects are not in competition.
Artifact: Student–faculty link
Represented the class throughout college: carrying concerns to faculty, bringing decisions back clearly, and learning how much patient communication sits behind a functional group.
Lesson — Representation starts with listening, then closing the loop.
Artifact: Campus coordination
Coordinated placement activity with students, faculty, and visiting companies while also helping the Computer Science Engineering Association run departmental events, workshops, and technical programmes.
Lesson — Good coordination removes uncertainty before it becomes friction.
Artifact: Peer mentorship programme
Founded and coordinated a Computer Science mentorship programme that paired junior teams with senior mentors for projects, placements, and the parts of college that are easier to navigate with someone beside you.
Lesson — Useful systems can be social systems too. Make the path easier for the next person.
Artifact: Degree
Graduated from Kongu Engineering College holding first rank in Computer Science, then went full-time at Soliton as Project Engineer.
Lesson — Fundamentals compound. So do side projects. Keep both.
Chapter 03 · 2021 – 2023
Medical-imaging and wildfire models — and the public datasets underneath them.
Artifact: Public dataset
My first artifact that left my hands and kept going without me — 24,000 balanced OCT scans, now indexed in the openmedlab Awesome-Medical-Dataset collection.
Lesson — A benchmark’s value is its fairness, not its size.
Artifact: Conference paper
Retinal disorders develop slowly and without obvious signs. This paper — and the 24,000-image OCT-C8 dataset behind it — tried to make them visible earlier.
Lesson — Infrastructure outlives results. The dataset travelled further than the paper.
Artifact: Public dataset
130,000 images, 8 cancer types, 26 subclasses, normalised into one corpus and published openly. It has since seeded well over a hundred community notebooks.
Lesson — The unglamorous work — renaming, resizing, balancing — is what other people actually reuse.
Artifact: Journal article
A sixth-semester classifier project matured into open-access Fire Ecology research: Xception reached 98.72% on the original dataset, and Learning without Forgetting kept 96.89% of it while learning an unseen one.
Lesson — Research is stronger when the dataset, demo, and limitations are all visible.
Artifact: Journal article
Eight cancer types, one set of models, and Learning without Forgetting so the network keeps what it already knew. Published in IEEE Access, volume 11.
Lesson — A model that forgets is a liability, not a curiosity.
Chapter 04 · 2021 – 2023
First production work: frontend UX, tooling, and services people had to live in.
Artifact: Changelog
Internship at Soliton Technologies: shipped a responsive dark theme, a drag-and-drop scheduling UX, and a CLI for multi-device monitoring.
Lesson — Real users change how you build. Immediately.
Chapter 05 · 2023 →
Ecommerce frontend at scale — React product experiences, accessibility, and experiments.
Artifact: Badge
From services to product. React product-page experiences for a global ecommerce customer base — accessibility, performance, and experimentation-driven delivery.
Lesson — At scale, small frontend details are big business.
Artifact: Role change
From Associate to owning ecommerce features end to end — technical discovery through production, alongside Product, UX, QA, Backend, and Analytics.
Lesson — Ownership is mostly communication with a deadline attached.
Chapter 06 · 2025 →
Shipping small, useful products end to end — and rebuilding this site as the workbench.
Artifact: theflames.app
One of the first hobby projects I wanted to build finally reached a polished public version: auto reveal, manual notebook mode, anonymous charts, and shareable results.
Lesson — Playful software still deserves production-grade care.
Artifact: shadyside.app
A 2020 Jalakandapuram-to-Coimbatore bus ride made the problem painfully obvious. The beta stayed small, but 1,000+ shade recommendations later it grew into ShadySide v1.5.
Lesson — A niche idea can still be real if the frustration is sharp enough.
Artifact: This website
The portfolio rebuilt as an interactive workbench — every module on the bench is a live door to real work, and every claim on it links to a source.
Lesson — Your own site should be your most honest product.
Chapter 07 · No end date
Practices with no ship date, kept because they change how I look at everything else.
ONGOING-01
A camera changes how you look at light — useful training for someone who ended up building a sun-tracking app. FrameOS is the work-in-progress home for the archive.
Lesson — The lens sharpens the angle you take on everything else.
FrameOS repo ↗ONGOING-02
Slow feedback loops, honest failure states, and no hotfixes. The best available counterweight to a career spent shipping software that changes hourly.
Lesson — Growth takes the time it takes — plants and code both.
The log stays open.
New entries land when there is something real to add — usually a shipped thing, occasionally a lesson worth keeping.