PROJECTS

Things I've built, end to end.

Apps real people use, systems work that goes down to the machine code, and an independent research paper, each one shipped and documented. Open any project for the full story.

012026Actively maintained

ShreyaDesk AI

Indian equity research, with execution guardrails

ShreyaDesk AI brings Indian equity research, technical scans, paper trading, and guarded live execution into one local dashboard. A React/Vite frontend works with a Python/FastAPI backend, cached NSE daily history, and Zerodha Kite Connect. OCC, RSI, Supertrend, SMA, and confirmed 3D workflows support research; account-ownership checks, persistent loss limits, order caps, and a kill switch guard execution. Live trading starts paused and must be explicitly enabled.

Case study ↗Project brief

022026Actively maintained

Satellite Crop Health Scanner

Pixel-level crop health from Sentinel-2 imagery

Satellite Crop Health Scanner converts aligned Sentinel-2 Level-2A B04 and B08 GeoTIFF bands into pixel-level NDVI analysis. It validates the full raster grid, handles invalid pixels safely, classifies vegetation with adjustable thresholds, and presents maps, distributions, and summary statistics. Results can be exported as a presentation-ready PNG, statistics CSV, or georeferenced NDVI GeoTIFF.

2Sentinel-2 bands3export formats

Case study ↗Live ◉ Source

crop-health-scanner.streamlit.app

032024Actively maintained

FinancialFlow

Income, budgets, and goals in one clear view

FinancialFlow brings income, expenses, category budgets, and savings goals into one personal finance dashboard. Transactions update balances and monthly summaries, while interactive line, bar, doughnut, and radar charts reveal spending patterns and financial progress. Built with Next.js and TypeScript, it saves data in browser local storage so users can return to their records without creating an account.

Case study ↗Project brief