Utkarsh Dixit
Full stack software developer in Kanpur, India. I build web apps, Android tools and the APIs behind them. Lately that means fraud detection at Frautect, and tools that track down fake accounts.
- Now
- Full Stack Developer at Frautect
- Local time
Look-alike finder
The copycat check from my Insta Accounts app, running in your browser. Type any username.
Accounts it would search for, and how suspicious each one looks
Recent work
Newest first. A lot of this is private company or client code, so those projects are explained here instead of linked.
-
2026
Insta Accounts
Android app that finds copycat and impersonator accounts on Instagram and exports them to Excel.- You enter a brand or a name. The app drives Instagram through an Android accessibility service: it searches each term, opens the Accounts tab, scrolls, and collects every result.
- It can also search look-alike spellings, like the ones in the finder above, and flags each account it finds: look-alike username, risky words such as “support”, numbers standing in for letters, not verified.
- Saves an Excel file with each account’s profile photo, plus an HTML copy for phones without a spreadsheet app. The .xlsx writer is plain Java with no libraries.
- Backs off by itself: it pauses 30 to 60 seconds every 10 searches and stops after 5 failures in a row, which usually means Instagram is rate-limiting.
- Signed APKs are built by GitHub Actions, so a new version installs straight over the old one.
Read the write-upTry the look-alike checkPrivate repository
-
2026
Telegram Keyword Search
Finds every message in a Telegram group or channel that mentions a keyword, with a direct link to each one.- A password-protected web app and a command-line version for Windows, macOS and Linux.
- Results appear as they are found and export to a CSV that opens correctly in Excel. A long search can be stopped without losing what it found.
- Two modes: Telegram’s fast server-side search, or a deep scan that reads every message to catch the keyword inside longer words.
- Works with several Telegram accounts. Sessions stay in the user’s browser; the server keeps nothing on disk, so redeploys lose nothing.
- Deploys to Render’s free plan in one click from a Dockerfile.
-
2025
Frautect mobile app
React Native app that scans the apps on an Android phone as part of Frautect’s fraud detection.- Built with Expo and a native module that lists every installed app with its package name and version, for a quick apps check and a deeper scan.
- Connects to Frautect’s cloud APIs so detection happens on the phone the user is holding.
- The native module became its own public npm package, listed next.
Company project at Frautect
-
2025
react-native-installed-apps-android
npm package that gives React Native apps the list of user-facing apps installed on an Android phone.- A Turbo Module for React Native’s new architecture, with TypeScript types.
- Returns only apps that have a launcher icon, so background system services don’t clutter the list, and includes each app’s version.
- Three major versions published so far.
-
2025
WalletWise
Expense tracker for Android and iOS with sign-in, income and spending totals, and a REST backend.- React Native app built with Expo Router, with authentication by Clerk.
- Express API on serverless Postgres (Neon), rate-limited with Upstash Redis.
- A scheduled job keeps the free-tier server awake so the app never waits on a cold start.
-
2025
Fraud detection dashboard
Web dashboard for logging fraud reports, running investigations and keeping a blocklist.- Reports and investigations each carry a status, and confirmed entities go on a blocklist that records who blocked them and when.
- Next.js with Prisma on MongoDB, sign-in by Clerk, and webhooks to keep users in sync.
-
2025
EventConnect
Marketplace for booking event services: browse providers by category, book them, and leave reviews.- Search providers, filter by category, book one, and follow the booking’s status.
- Next.js with a MongoDB API for providers, categories, bookings and reviews, validated with Zod.
Private repository
-
2024
Kaarigar-Sathi
Connects people with local tradespeople, with live tracking of each job.- Real-time job tracking so customers can see where a job stands, used by around 1,200 people a month.
- Location-based matching raised successful connections between customers and tradespeople by 40%.
- Node.js and Express API on MongoDB, built for heavy traffic.
-
2023
PriceWise
Paste a product link and get an email when its price drops.- Tracks prices for 500+ users across more than 10,000 product records.
- Welcome emails and price alerts sent with Nodemailer.
Earlier projects
| Year | Project | Built with | Link |
|---|---|---|---|
| 2024 | News-Eve, a daily news reader with categories | React, News API | Code |
| 2023 | SnapWorld, a photo-sharing social app | React, Node.js, MongoDB | Code |
| 2023 | PinCraft, a Pinterest-style pinning app | Express, MongoDB | Code |
| 2023 | Amazon storefront clone with Stripe checkout | React, Firebase, Stripe | Live |
Where I’ve worked
-
to now
Full Stack Developer
- Building Frautect’s fraud detection platform end to end: the web dashboard, the APIs, and the React Native app that scans phones.
- Deploying and scaling services on AWS with CI/CD pipelines.
- Reviewing code and setting coding standards with the team.
-
to
Full Stack Developer Intern
- Built the client dashboard in Next.js and MongoDB: a multi-step company profile form, a real-time threat monitoring page, and an overview with interactive charts.
- Wrote REST APIs with concurrency control, so several people from one company can edit the same records without overwriting each other.
- Published the react-native-installed-apps-android package and built the first version of the mobile app.
-
to
Full Stack Developer Intern
Madrocket Technologies (iQNaut)
- Built a role-based school ERP on Node.js, Firebase and Firestore for more than 1,000 students and staff.
- Cut API response times by 40% by reworking queries and the Firestore schema.
- Tested with varied datasets and fixed over 100 critical issues before launch.
About
I’m Utkarsh Dixit, a software developer based in Kanpur, Uttar Pradesh. I studied computer science at Dr. A.P.J. Abdul Kalam Technical University in Lucknow, graduating in 2024, and I’ve been shipping production code since my first internship.
I like the whole stack: a database schema that holds up, an API that doesn’t fall over when two people click at once, and an interface people understand without a manual. Most of my recent work sits where security meets everyday users, spotting fraud, fake accounts and risky apps.
Outside work I solve algorithm problems. I’m past 800 on LeetCode.
What I work with
- Languages
- JavaScript, TypeScript, Java, Python, SQL
- Web
- React, Next.js, Tailwind CSS
- Backend
- Node.js, Express, FastAPI, REST APIs
- Data
- MongoDB, PostgreSQL, Prisma, Firebase, Redis
- Mobile
- React Native, Expo, native Android
- Shipping
- AWS, Docker, GitHub Actions, Cloudflare, Render, Vercel
Recognition
- 800+ LeetCode problems and 20+ badges. See profile
- Top 20 of 150+ teams at Hack This Fall 2024, Karnavati University
- Top 10 of 100+ teams at Code for Good, Delhi Technological University
- Java certification from HackerRank. See certificate
Say hello
I’m open to full-time roles and freelance projects, remote or in person. Email is the fastest way to reach me.