Bridging four years of high-performance mobile architecture with autonomous machine learning.
Pragmatic, battle-tested client architecture built from the ground up—shipping live ride-hailing dispatch, crypto-enabled fintech rails, and real-time logistics systems.
Led frontend architecture for a ride-booking application operating exclusively in Srinagar, India. Built the app from the ground up, managed deployments, and tracked live reporting.
Integrated proprietary BlockCoin cryptocurrency directly into the app. Implemented BBPS services for 35+ utility bill payments and built an MLM commission system.
Core screens from the BlockPay build — wallet, BBPS rails, commission engine & settlement bridge.
Built full-stack solutions via UpWork including ParcelApp (Google Maps API, live Firebase tracking, Portuguese localization). Previously taught myself Java/XML, published 5 educational apps for HSC students, and built a Kotlin/Firebase system for a local laundry business.
ParcelApp (UpWork)
Google Maps API, live Firebase tracking & Portuguese (pt-BR/pt-PT) localization.
5 HSC Educational Apps
Self-taught native Java/XML Android engineering shipped to students.
Laundry Operations System
End-to-end Kotlin + Firebase order & dispatch workflow for local business.
Architecting autonomous multi-agent workflows, low-latency neuro-telemetry pipelines, and first-principles deep learning models.
Currently building intelligent, autonomous AI agents for university-wide deployment, bridging my heavy software engineering background with advanced machine learning.
Engineered a core BLE communication layer connecting a hardware headband to read live EEG and tDCS brain signals. Plotted real-time data onto a doctor-facing interface and wrote Python scripts to stress-test the data pipeline.
Engineered a Convolutional Neural Network (CNN) entirely from scratch for hand sign recognition without relying on 3rd party black-box models. Led the core technical architecture for a team of 4.
Architected custom forward/backward spatial hierarchies & led pipeline integration across 4 researchers.
Shipping agents and reproducible analytics in the open — from Gemini-powered assistants to deployed data apps.
An agentic AI assistant built on Google Gemini 2.5 — streaming LLM responses, multi-tool data routing, and a cross-platform Flutter deployment.
End-to-end data science projects: real Kaggle datasets turned into interactive Streamlit apps across the full workflow — exploration, modelling, and deployment.
Exploratory analysis and modelling of equity time-series — ingesting market data, engineering features, and surfacing trends through reproducible notebooks.