Harsh Gopal
Builder of AI-powered systems and applications.
I'm Harsh Gopal, a Computer Science Engineer focused on
building software at the intersection of AI, systems and products.
Core Capabilities
I like to build real software and AI systems.
From AI agents and real-time applications to experimental systems, while focusing on turning ideas into working products.
This portfolio is itself an example of that approach — a digital system designed around interaction, architecture and technical storytelling.
Cart Radar
How can product availability across quick-commerce platforms be aggregated across a geographic radius in real time without official APIs, without getting blocked?
A real-time inventory aggregation platform that tracks product availability across Blinkit, Zepto, Swiggy Instamart and BigBasket. Uses hexagonal grid sweeping, Playwright API interception, asyncio semaphores, and SSE streaming for live results.
MySphere
How can multiple physical devices be made to behave as a single coherent platform — sharing state, clipboard, and control without fragmented apps?
A Flutter-based cross-device ecosystem that treats multiple devices as one unified platform. Syncs clipboard, controls media, and supports file transfer across macOS, Windows, and Android via local WebSocket and Firebase cloud relay.
Adaptive 3D GameplayIEEE RESEARCH
How can game difficulty adapt dynamically to a player's skill level during a session to maintain flow state — without manual tuning?
IEEE-published research exploring how adaptive AI systems improve player engagement in 3D game environments by dynamically adjusting difficulty based on real-time player performance telemetry.
XYZ Stocks
How can retail investors in India access institutional-grade AI stock analysis without expensive subscriptions or technical expertise?
An AI-driven investment intelligence platform built for the Indian market. Analyzes NSE/BSE stocks using live market data, Gemini AI insights, IPO tracking, momentum analysis, and portfolio health scoring.
FamilyVault
How can families securely share, organize, and access files across devices without relying on third-party cloud services — maintaining full control of their data?
An offline-first, multi-tenant file sharing system with RBAC, group management, and device pairing. Ships as a native macOS Electron desktop app and Android Jetpack Compose mobile app, with a Go backend powering the core.
I experiment with conceptual AI architectures and intelligent workflows before they hit production.
Where I've worked
Independent Software Engineer / Builder
INDEPENDENT- ›Building production-oriented AI and full-stack systems independently.
- ›Developing Cart Radar, a real-time quick-commerce inventory aggregator using Playwright and SSE.
- ›Building XYZ Stocks, an AI-powered stock analysis platform leveraging Gemini and live market data.
- ›Exploring cross-device ecosystems (MySphere) and agentic workflows (MCP, LangGraph).
AI/ML Intern
EDUNET FOUNDATION(Microsoft & SAP CSR)- ›Built shopping trend analytics prototypes using Python and ML workflows.
- ›Worked on data preprocessing pipelines for structured and unstructured datasets.
- ›Collaborated with teams to validate ML model outputs and improve prediction accuracy.
Data Engineering Intern
EMULUS CONSULTING- ›Performed data cleaning and validation on large datasets.
- ›Produced technical documentation for data engineering workflows.
- ›Contributed to improving data pipeline reliability and consistency.
Let's build
something.
I showed you what I build.
Now let's build something together.