Aryan ChopraProduct · Engineering · AI
BLR
Aryan ChopraThe colonnade

Hanging the paintings.

AryanChopra

Product person first, engineer and builder at heart. I’m building the next generation of the creator economy, with products 200,000+ people already use.

Now
Building the next generation of the creator economy
Focus
Consumer products, AI agents, and the systems that keep them fast and reliable
Based
Bengaluru, India

About

Product first, builder at heart.

I'm a product person first, and an engineer and builder at heart. In college I ran two technical clubs and grew the community to 500 people. On nights and weekends I shipped things for strangers: a browser agent that reads the page, a coding assistant with 184 stars on GitHub, and a community that about 10,000 people joined.

Since then I've taught a model to do the books for 10+ companies at Dunlin.AI and built analytics for Zerodha's Zero1 network. Now I'm on the founding team at WLDD, where Solo and Memed have grown past 200,000 users. I care about how people actually use things, and I like building them end to end.

Off the clock, I work out every day and analyse my own health data. Yes, I’m a nerd.

Study in dots, drawn in code

Experience

The work, so far.

WLDD

Founding team · Backend leadAug 2025 — Now

I own the backend behind Solo and Memed, two consumer apps with 200,000+ people on them: money movement, real-time messaging, data pipelines and the AI layer, end to end. I also lead the developers and interns who build on it.

  • Payments
  • Real-time
  • Data pipelines
  • LLM agents
  • Team lead
200k+people on the systems I run
  • Real-time messaging between creators and brands over WebSockets, built into Solo.
  • Brand intelligence for Spotify, Netflix and Rapido: a pipeline that scrapes Instagram and runs date-filtered search, then turns the noise into categorised audience sentiment.
  • Payments where a rupee can’t go missing: wallets, withdrawals and vendor payouts on Cashfree.
  • An LLM agent for our own dashboards: say what you want, and tool-calling turns it into real backend actions.
  • Leading the backend team: setting the architecture, reviewing the work and keeping frontend and backend shipping in step.
Built with
  • Node.js
  • MongoDB
  • Cashfree
  • WebSockets
  • LLM tool-calling
solocreators.appMemed on the App Store

Imagined Studio × Zerodha

Forward-deployed engineerFeb — Aug 2025

Embedded inside Zerodha's Zero1 media team. I took a sprawling YouTube network and made it one source of truth, then made years of long-form video searchable in plain English.

  • Analytics
  • Ingestion pipelines
  • Semantic search
  • Applied LLMs
6+channels unified into one view
  • Zero1 Intelligence: views, demographics, geography, devices and comment sentiment across GrowthX, Think School, Finshots TV and more, in one dashboard.
  • An ingestion pipeline for long-form video: transcripts split wherever the script changes topic, and Gemini writes a structured summary of every chunk.
  • Semantic search over the whole archive: chunk embeddings in ChromaDB, so a plain-English question lands on the exact moment, with notes on what content performs.
Built with
  • Gemini
  • ChromaDB
  • Vector search
  • Next.js
Zero1 Intelligence

Dunlin.AI

Full-stack engineerAug 2024 — Jan 2025

An AI bookkeeping product. I built the engine that categorises every transaction for 10+ companies: one fine-tuned model, grounded in each company’s own history, fast enough to feel instant.

  • Multi-tenant AI
  • RAG
  • Fine-tuning
  • Integrations
<5sto categorise 400+ transactions
  • Multi-tenant by design: one shared fine-tuned Azure OpenAI model serves 10+ companies, with each company’s context injected per request instead of a model per client.
  • Grounded in real history: 100k+ past transactions embedded in ChromaDB, self-hosted on AWS. The closest matches go to the model as examples, so every category has evidence behind it.
  • Integrations that remove the busywork: a Manifest V3 Chrome extension that securely captures encrypted QuickBooks and Zoho sessions for automatic sync, plus REST APIs for both. Onboarding got twice as fast.
  • A Next.js dashboard for reviewing and correcting 1,000+ transactions.
Built with
  • Azure OpenAI
  • Fine-tuning
  • ChromaDB
  • AWS
  • Chrome extension
  • Next.js
dunlin.ai

Projects

Built on nights and weekends.

MyPeachA safe place to talk about what nobody talks about.A sexual-wellness community for India that grew to about 10,000 members. I built all of it, end to end: login, posts, profiles and a reels-style feed.Next.js · Postgres~10k membersCrack CoderThe cheat code for coding interviews.A desktop app that sits beside browser coding tests, out of sight. It reads the question, runs your code and finds the answer with AI.Electron · TypeScript184 starsAutoSurfComputer use, before it had a name.Give it a task in plain English and it takes the wheel: it opens a browser, reads the screen with a vision model, and clicks and types on its own until the job is done.Node.js · Puppeteer · OpenAI visionAutonomous agentStandup AIToo bored to attend standups, so I built something that goes for me.Agents that join the standup on my behalf: they give my update, take the notes and message me what I missed.Multi-agent LLMOpen source

Contact

Building something
worth using?

If you're a small team building something people will actually use, I'd like to hear about it. Email is fastest.