Aryan Singh Sengar

Builder — 17

Profile

17-year-old who builds software and likes solving hard problems that look like science fiction today. I work under cost constraints so that what we make stays affordable for both makers and users, not a luxury. Right now I want to test low-cost wireless EV charging through highway guardrails.

Projects

Maglion · simulated, grant applied for to build the prototype

Founder

  • Wireless EV charging built into highway guardrails and median barriers, with charging pads for cities, powered by solar panels along the roadside.
  • Goal: smaller EV batteries and less charging time. Not yet built or measured; the first step is a low-cost bench test.
  • Why it matters: the battery is the biggest cost of an EV. If cars charge while driving, they could carry a battery up to about half the size. This is my estimate from simulations, not a measured result.
  • Dynamic wireless charging has been done by an Israel-based company, but set-up and maintenance cost is the problem; Maglion aims to cut it by 3× to 5× using guardrails and solar power (a target, not yet proven).
  • Computer simulations show Maglion cutting battery, operating and building (infrastructure) costs.
  • Recently applied to Emergent Ventures for a grant to build the prototype; awaiting a reply.

Agentidin · hybrid agentic professional network, in development

Founder

  • Started after struggling to hire people with the right skills for Wynko.
  • Every person has their own agent: brought from elsewhere (e.g. OpenClaw), connected through AI tools like Cursor, ChatGPT or Claude, or made on the platform. The agent builds the person's profile from GitHub projects, Codeforces and verified skills, so they are represented even when offline.
  • Each company has an agent connected to its company brain and Slack. With a Claude tag or a command, the company describes the skills it needs and its agent finds matching people's agents.
  • Agents talk to agents, agents to humans, and humans to humans, without the cost of a LinkedIn recruiter seat that early-stage startups and companies in developing economies struggle to afford. Aim: cut hiring time sharply while staying human-friendly.

Wynko · software

wynko.in

  • Windows desktop app, Android app, Focus Lock browser extension and website, built on a low-level kernel system, that help a person stay away from distractions while studying, giving breaks and blocking autonomously.
  • Recruited a team from Reddit to build it alongside me.
  • Founded and ran an agency with a team of 4 to fund Wynko's development: website development, tools and social media handling for clients, including shopkeepers. The capital it generated paid for our development tools and stipends for the team.

Low-cost AI coding setup · built when I could not afford paid tools

Tooling

  • Hosted open-source models on free cloud notebooks such as Kaggle, with Qwen as the text model.
  • Added LiteLLM as the gateway, Llama on Groq's free tier as the vision model, and a headless browser. The text model writes the code and the vision model checks how the page renders.
  • Wrapped it all as one API endpoint and used it inside Claude Code.

Kinux · cross-embodiment robot learning, research in progress

Research

  • Building an SDK for robot failure logging that lets data move from one robot type to another with maximum positive transfer.
  • Idea: classify robots the way Carl Linnaeus classified living species, grouping them by how much one robot's parts (motors, joints, torque) transfer positively to another, so a robot can learn from the best-matching donor type.
  • Early result: randomised combinations of parts from multiple robots, matched which parts give the most positive transfer, and stress-tested the combined result. So far this beats transferring a whole robot to another whole robot; early results that may change with more testing.
  • Status: still researching classification by maximum positive transfer, identifying failures, and 3D AI simulation for learning from other robots.

Skills

Languages
Java, Python, Rust, Kotlin, HTML, CSS
Platforms
Linux, Android app development, Windows app development
Web & design
Next.js, Tailwind CSS, Figma
AI
Orchestrating multiple AI agents at a time
Learning
I learn any technology fastest by building with it rather than reading theory.

Background

IISER IAT
Rank 3846 in the IISER Aptitude Test, the entrance exam for India's premier research institutes: the IISERs, and also used by IISc and IIT Madras for their BS programmes.
Self-taught
Taught myself to code in 4th standard with no computer at home, writing programs in a notebook and running them on school computers; wrote 200 to 300 programs a day. Got my own computer only after 10th standard.
Debating
Won several debate competitions in school; I love debating.
Home
Khajuraho, a small town in India.