Prashant Ghimire

I'm a software engineer based in Kathmandu, Nepal. I've been doing Flutter for 5+ years, and for the last 2 of those a lot of my work has been inside agentic dev workflows.

My first real Flutter role was at Ramailo Tech, working remotely with their Bangalore team. The main project there was a digital gold purchasing app powered by Augmont (India's largest gold platform), plus an internal logging tool for Physics Wallah's coaching classes, with weekly delivery to a US-based client. After that I joined BitpointX (remote, based in Sydney) and designed and built a Flutter app for an Australian driving instructor and learner marketplace, with geolocation-based booking and in-app payments. At EKbana Solutions in Lalitpur (onsite) I led the Flutter team, ran code reviews, and mentored juniors on SOLID and clean architecture. I shipped a production digital wallet there (money transfer and utility bill payments) and a live trivia game with thousands of concurrent players. I also owned the R&D phase before greenfield builds: Bluetooth Low Energy, on-device object detection with PyTorch Mobile and TensorFlow Lite, OpenStreetMap tile caching, and Siri intents via App Intents. Currently I'm at Xuno on a remote contract, mostly shipping Flutter features end to end inside an AI-augmented dev workflow with Claude Code. Some Python CRUD endpoint work and internal data tooling too, and I contribute to solution architecture when features are in planning.

Outside of work I publish indie Android apps to Google Play. To do that sustainably I built a personal multi-agent Claude Code framework that handles the boring parts: scaffolding new apps, syncing analytics events, generating compliance pages, and shipping updates. It uses role-specific subagents, PostToolUse hooks, and cross-app Play API automation. I also have a couple of Flutter packages on pub.dev (interpolated_countdown for color-interpolated countdown widgets, dynamicform-flutter for declarative multi-page forms) and a few open-source Python projects, including Autoreels (a near-fully-automated reel-video pipeline) and MASH (a multi-agent healthcare simulation).

The things I find most interesting right now are agentic engineering at production scale, on-device ML for mobile, and the blend of fast indie shipping with proper R&D. I'm comfortable end to end, so I tend to do my best work in roles that let me span mobile and agentic AI tooling rather than picking one.

Based in Kathmandu (UTC+5:45). Flexible with working hours and happy to align with your team's schedule.