Flutter Developer Intern (Mobile and AI Video)

Nada is India's first social network for arranged marriages. Matches come through the people a family already knows: relatives, friends and the community matchmaker, a role the app calls Community Member. The app syncs a user's phone contacts, builds a who-knows-whom graph, and shows how two profiles connect. It has been live on the Play Store and the App Store since August 2026, and the full launch is on 11 October 2026. Nada Network Private Limited is a DPIIT-recognised deep tech startup working from Noida.

The founder, Harsh Titoria, built the backend and runs product and engineering end to end, including the app. He was a backend engineer at Dell Technologies before Nada, after a BTech and MTech from IIIT Gwalior. You would be the second person writing code for the app.

• Flutter 3.44 and Dart 3.10, one codebase for Android and iOS, live on both stores.
• Riverpod 3 with the Notifier pattern for state. Screen to provider to repository to model, and never backwards.
• Dio with a JWT interceptor and automatic token refresh, WebSockets for chat, Firebase for phone authentication, Crashlytics and Analytics.
• Plain Dart model classes, with no code generation (no freezed or json_serializable).
• A custom design system: tokens for colour, type, spacing and motion, with Fraunces and Manrope.
• Widget tests, golden tests with an injected clock, responsive tests across compact phone widths, and dart analyze at zero issues before every commit.
• About a dozen feature modules: onboarding, discover, the connection chain, requests, Community Member mode, chat, notifications, rewards.

Flutter engineering on a live app, with the founder reviewing every pull request.

• Build screens and flows. There is a written list of screens for the next app update, across onboarding, profiles, discovery and chat, each with a spec and the backend contract. You take a screen from its spec to a merged pull request: the models, the provider, the repository call, the screen and the tests.
• Fix bugs. Issues from the store review queue and crash reports: a keyboard covering an input on a small phone, an empty state, a stale cache. Each fixed with a widget test so it cannot come back.
• Take on new requirements. The launch brings users, and the requirements change weekly. You scope them with the founder and build them.
• Both platforms. An Android device or emulator is enough to start. iOS builds need a Mac. If you have one, you run both. If not, the founder runs the iOS side.
• Make AI videos for social media. The person who runs Nada's channels sets the brief, you write the prompts, generate and pick the takes, cut them to length with subtitles, and deliver in the sizes each platform needs. The founder signs off before anything goes out. This runs alongside the Flutter work through the launch season.

Every pull request is reviewed by the founder, over screen-share for the first month.

Nada pays for an AI coding assistant for you, Claude Pro or similar, on a company account, and for an AI video account, Higgsfield or OpenArt, for the video work. Use them for the work. Screens from a written spec are where coding assistants do best. Three rules: you can explain every line the assistant wrote before you open a pull request, nothing from production and no user data ever goes into a prompt or a video, and the accounts are closed on your last day.

• You have shipped a Flutter app, or built one substantial enough to show, and your GitHub or a store link shows it.
• You write Dart with a state management library, Riverpod or another, and you can explain why the layers are separated.
• You have integrated a REST API with authentication and handled the loading, empty and error states.
• You have written at least a few widget tests and want to write more.
• You have generated video or images with AI tools, Higgsfield, OpenArt, Runway, Kling, Veo or similar, and you can show pieces with the prompts behind them, one of them a video if you have had access to a video tool.
• Git and pull requests.
• You have 24 hours a week, Monday to Saturday, about four hours a day, for three months, and a laptop that runs Flutter.
• You are a student or a recent graduate.

Good to have. Golden tests. Firebase. WebSockets. Accessibility work. A Mac with Xcode. An app of yours on a store. A reel you made with AI video and cut yourself.

• ₹5,000 a month, an internship letter, and a certificate and a reference at the end.
• A paid AI coding assistant and an AI video account, Higgsfield or OpenArt, on company plans.
• Code review on every pull request from the founder, who runs the app and the backend.
• Your screens in an app that is live on both stores.

1. Apply with the form, your GitHub profile or a store link, and pieces you generated with AI tools, videos or still images, with the prompt behind each. We look at your projects first.
2. Shortlisted candidates get a task by email: a small two-screen Flutter app from a spec we send, with one widget test. You have two days from the email to send it back as a GitHub repository link with a README that says how to run it. We run it and read the code.
3. If the task is selected, an interview with a member of the team about how we work.
4. A final interview with the founder, starting from your task.

Remote, with the camera on for calls. A short written check-in every working day and a short team call on working days. Company accounts for every tool.

Nada hires across gender, religion, community and background.