Work — 09 projects

Work

Every project below is live code: real screenshots, public repositories and, where they exist, working demos. Open any project for screenshots, the full write-up and its live link.

J.A.R.V.I.S — NIX application screenshot
01

J.A.R.V.I.S — NIX

Cloud assistants are subscription-locked, can't touch your OS, and forget everything between sessions.

PythonGemini Live APIPlaywright
15+assistant capabilities
LumoroX Park application screenshot
02

LumoroX Park

City parking is scarce while thousands of private driveways sit empty all day; the original system proved slot allocation but needed a real marketplace around it.

TypeScriptReactSupabase
LiveOpen live site
LumoroX application screenshot
03

LumoroX

The original collaborative filter proved personalized top-10s, but users still needed a modern surface for mood, search, trending and watchlists.

TypeScriptReactSupabase
LiveOpen live site
Sophia — AI Assistant application screenshot
04

Sophia — AI Assistant

Existing assistants are shallow — no persistent context, no multi-step task planning.

PythonLLMsPrompt Engineering
Multi-stepplanning and recall
Telecom Customer Churn Prediction application screenshot
05

Telecom Customer Churn Prediction

Telecom operators lose ~15–25% of subscribers yearly, with no early-warning signal.

PythonScikit-learnXGBoost
87%model accuracy
Movie Recommendation System application screenshot
06

Movie Recommendation System

Cold-catalog discovery: helping users find relevant titles beyond top-N popularity.

PythonSurpriseSVD
18%RMSE improvement
Parking Space Management System application screenshot
07

Parking Space Management System

Urban lots waste ~30% capacity to poor visibility & manual allocation.

PythonSQLStreamlit
IJCRTpublished research
Table Tales application screenshot
08

Table Tales

Small restaurants juggle paper menus, phone orders and walk-in table bookings with no single place to manage them.

JavaScriptFirebaseAndroid
Real-timeOpen live site
AI Developer

Generative AI Features

LLM-driven flows embedded into Android & web apps with prompt orchestration.

Prompt Engineering

Structured templates, evals, and guardrails for production assistants.

Applied ML

Classification, regression, and recommender systems on real datasets.

Agents & RAG

Currently exploring agents, retrieval-augmented generation, and on-device inference.

GitHub

Loading stats…