Anirudh Vasudevan is a full-stack and AI engineer based in San Francisco. He is a Member of Technical Staff at Rivo, a San Francisco fintech, where he builds production voice AI agents and owns the analytics and growth stack. He has won five hackathons, holds a granted patent for a contactless voting system, and co-wrote a paper accepted to EMNLP.
This page is about the software engineer. There are other people with the same name, so if you are looking for the engineer behind LARK, Living Photos, or ScreenSense, you are in the right place.
What I do now
I am a Member of Technical Staff at Rivo. Most of my week goes into conversational AI that real customers talk to:
- Voice and agentic AI. I deploy conversational and agentic AI in Python on cloud microservices, including LiveKit voice agents that capture leads and an internal support chatbot.
- One app, two platforms. I own the iOS and web apps, which run on a single React Native codebase that I migrated from Next.js. Playwright and Maestro end-to-end tests and Sentry monitoring keep it honest.
- Speed. I added TanStack Query caching with endpoint-level invalidation, which cut homepage load time by 87 percent and improved p95 latency.
- Growth. I run full-funnel tracking and paid ads across TikTok, Meta, Reddit, and Google, built an admin-configurable referral system, and built a pipeline that generates UGC-style video ads end to end.
Before Rivo I was the founding engineer at Nonlinear, a San Francisco AI startup. There I built the Universal Node System for AI workflows, which cut configuration time by 45 percent, and agentic RAG pipelines behind context-aware agents and a Microsoft Teams bot, which cut retrieval latency by 60 percent.
Before that I was a graduate research assistant at the University of Minnesota's Institute for Health Informatics, doing data processing, causal analysis, and predictive modeling on K-12 education and health equity in Minnesota.
Hackathons
I like building a whole product in a weekend and putting it in front of judges. Five of those weekends ended with a win or a podium finish:
| Result | Event | Project |
|---|---|---|
| Top 10 | Y Combinator AI Hackathon | LARK, a multi-agent MCP gateway that turns ChatGPT and Claude into a phone |
| Winner | ElevenHacks, Stripe x ElevenLabs | Living Photos, walk into an old photograph as a 3D scene |
| 2nd Place | ElevenLabs x Replit | FaceTime from Mars, voice calls with AI colonists on Mars |
| 3rd Place | ElevenHacks, ElevenLabs x Zed | Apartment 4B, a first-person horror game voiced with ElevenLabs |
| Best Voice Agent | ElevenLabs x Firecrawl | ScreenSense, a Chrome extension you control by talking |
Across ElevenHacks Season 1 I submitted a project to all 11 hacks and finished 4th overall on the season leaderboard.
I have also been invited to build at hackathons hosted at Y Combinator in San Francisco, including the Moss conversational AI hackathon, where I built Tax It Easy, a voice CPA that only says what it can back up with your tax return.
Things I have built
- RUA, a privacy-first app for family caregivers. I built the iOS app (TestFlight) and the web app solo, with an integrated voice agent, encrypted on-device storage, and paperwork automation that drafts insurance appeals and renewals. The technical story is in Building RUA offline-first, and the waitlist is at rua-care.netlify.app.
- Anirudh's World, a 3D village that lives on the real clock with no backend, because the whole world is a pure function of a seed and the current time. Clear the monster camps to unlock portals to my projects. Read how it works or play it.
- An automated YouTube Shorts pipeline that turns a topic into a published Short (LLM scripting, fal.ai images and video, ElevenLabs voice, Whisper subtitles) and runs four channels on a daily schedule with a review dashboard.
- This website, including the AniBot chat assistant and the static blog engine that renders these pages. See how AniBot works.
Every project is listed on the Projects page, and every write-up is on the Writing page.
More: my full hackathon record, the patent and papers, and the tools I use.
Patent and research
- Granted patent: Contactless Integrated Voting System (filed as A System and a Method for Contactless Voting, Indian application 202341031598, May 2023), assigned to SRM Institute of Science and Technology. It lets people vote using speech and hand gestures instead of touching a machine. Here is the engineering behind CIVS.
- EMNLP (Main Conference): ConsensusDrift: Semantic Salience Can Override Majority Evidence in LLM Aggregation, with Gokul Srinath Seetha Ram and Rashmi Elavazhagan.
- Paper: A Deep Convolutional Neural Network for Remote Life Activities Detection Using FMCW Radar Under Realistic Environments (Springer, 2023).
- More on Google Scholar and ORCID.
Education
- MS in Computer Science, University of Minnesota, Twin Cities (2023 to 2025)
- B.Tech in Computer Science and Engineering, SRM Institute of Science and Technology (2019 to 2023)
- Diploma in Data Science, IIT Madras (2021 to 2023)
I have been writing code since tenth grade and have solved more than 500 problems on LeetCode. I grew up in India and now live in the Bay Area.
FAQ
Who is Anirudh Vasudevan?
Anirudh Vasudevan is a full-stack and AI engineer in San Francisco. He is a Member of Technical Staff at Rivo, where he builds production voice AI agents, and before that he was the founding engineer at Nonlinear. He is a 5x hackathon winner, holds a granted patent, and has a paper at EMNLP.
What does Anirudh Vasudevan work on at Rivo?
At Rivo, Anirudh Vasudevan builds conversational and voice AI agents in Python, owns the iOS and web apps on one React Native codebase, and runs product analytics and growth, including tracking, paid ads, a referral system, and an automated video ad pipeline.
Which hackathons has Anirudh Vasudevan won?
Anirudh Vasudevan placed in the top 10 at the Y Combinator AI Hackathon with LARK, won ElevenHacks (Stripe x ElevenLabs) with Living Photos, took 2nd at ElevenLabs x Replit with FaceTime from Mars, 3rd at ElevenHacks (ElevenLabs x Zed) with Apartment 4B, and won Best Voice Agent at ElevenLabs x Firecrawl with ScreenSense.
What is Anirudh Vasudevan's patent?
Anirudh Vasudevan holds a granted patent for the Contactless Integrated Voting System (patent ID 202341031598, May 2023), a voting system that uses speech and hand-gesture input so voters never touch the machine.
Where did Anirudh Vasudevan study?
Anirudh Vasudevan has an MS in Computer Science from the University of Minnesota, Twin Cities, a B.Tech in Computer Science and Engineering from SRM Institute of Science and Technology, and a Diploma in Data Science from IIT Madras.
How can I contact Anirudh Vasudevan?
You can email Anirudh Vasudevan at anirudhvasudevan11@gmail.com, message him on LinkedIn, or use the contact form on this site.