Proof-of-work archive · Field notes from the making

Things I've
built, broken,
studied & shipped.

This is not a list of skills. It is a collection of actual work — the questions, imperfect systems, useful accidents, and evidence left behind.

See the work ↓
No. 01
A living collection
Selected work / 2024—2026

Made in the real world.

Four projects, many rabbit holes.
Click around — the proof is the point.

Traffic Sign Recognition cover diagramUniversity project / 2024
AI / Computer Vision2024

Traffic Sign Recognition

A CNN-based recognition system for 43 traffic-sign classes, built with the GTSRB dataset and focused on the recognition component of a university group project.

Open the evidence
AI Inference Tool cover diagramProduct / 2026
AI2026

AI Inference Tool

A practical interface for testing and understanding model inference.

Open the evidence
Shelf Life cover diagramResearch / 2026
Research / Data2026

Shelf Life

A public library knows what was checked out — never what people almost chose. Built the missing signal from 912K real 'want to read' shelvings, then trained a small recommender that favors books people actually finish.

Open the evidence View the notebook
Carbon-Aware Computing cover diagramSystems project / 2026
Sustainability / Systems2026

Carbon-Aware Computing

A compute scheduler that shifts flexible workloads toward cleaner energy windows instead of treating time as neutral.

Open the evidence View the notebook
The archive

Browse by
kind of work.

Experience / 2025—2026

Work that shipped
with other people.

01Dec 2025 — May 2026

Data Engineer Intern · Infosys

Owned an end-to-end reporting pipeline in Python and MySQL, with validation and quality-control checks from ingestion through transformation to dashboard delivery. Wrote optimised SQL, built Power BI dashboards, and served as Scrum Master for a five-member team.

View internship certificate
02Project work

Trader Behaviour Analysis · PrimeTrade AI

Investigated trader behaviour and sentiment signals, then built Power BI dashboards around profit margin, win rate, drawdown, and strategy-level performance.

About the makerstill curious →

I like making things
that explain themselves.

I'm a builder, researcher, and occasional over-documenter. I work across data, product, automation, and the blurry edges between them — usually starting with a question that refuses to stay small.

My favourite part is the bit after the launch: looking at what people actually do, what broke, and what the work is trying to become.