Hi everyone! I am a student developer and builder from Bangalore, India.
I work on software and hardware projects that try to make capable technology more practical: usable on limited hardware, affordable, privacy minded, and based on real constraints, not just demos.
I’m interested in applied AI, embedded systems, local-first software, automation, privacy tooling, and the systems that make tech accessible to people.
- Constructing useful things from first principles.
- Making AI work on limited hardware or unreliable internet
- Local-first and privacy-preserving systems
- Embedded devices, sensors and hardware hacking
- Solid software architecture, not just nice once prototypes
- Recognize the trade-offs in security, performance, cost and usability
1st Place — Qubic Hack the Future, EasyConnect Track QubicLink is a no-code tool for identity verification and community automation in the Qubic ecosystem.
It explored a practical way for Discord users to prove ownership of a Qubic wallet without having to deploy smart contracts. The user would place a uniquely specified bid order on QX, creating an on-chain signal that could be associated with their Discord identity.
Once verified, that connection might initiate community workflows such as:
- Assign discord roles
- Eligibility for rewards and airdrops
- Control of access
- Community engagement and reputation management
For the project, Qubic EasyConnect and Make.com were used to link the activity on the blockchain to existing community tools. The objective was to offer DAOs, launchpads and other Qubic-based communities wallet verification and engagement automation without the need to deploy custom smart-contracts.
Anonymous messaging Slack bot, built during Hack Club Stardance.
You can toss messages in a shared bottle read random messages later. The project was designed to keep messages simple and anonymous, so it only stores the message text and timestamp, not user identities.
I’ve also explored ways to make AI more accessible if you don’t have a “fast laptop + constant internet” setup:
- SMS-AI - a proof of concept to access AI via SMS and basic phones
- Octo AI - Local inference experiments with speech input, text to speech and local language models (hands free)
These projects influenced my ongoing interest in low-cost, private, and resource-efficient AI systems.# Now searching
- AI systems running on device and offline
- Embedded hardware and wearable interfaces
- Web tooling for privacy and security
- Tools for developers and automation
- Backend architecture and distributed systems
- Algorithmic problem solving in C++
I like to get past the first demo and deal with the harder parts: constraints, failure cases, debugging, integration, performance and whether or not a system is actually useful to someone.
Most of the repositories here are experiments, hackathon projects or work in progress. They are the things I was trying to learn and build at that time.