Anay Shekhar

Anay Shekhar

15, building intelligence

about

i'm a sophomore in high school interested in machine learning, especially how models represent their surroundings and predict what might happen next.

now

  • how models represent objects, space, and motion
  • how they learn the way an environment changes over time
  • how those representations can support prediction and decision-making

projects

orit

an autonomous company that simulates product performance in the market, built for ecommerce and local companies. won best use of exa at the zero human company hackathon in nyc.

felix

a multi-agent system converting natural language and computer vision into real-world computer actions. won best use of gcp at tillyhacks.

fern.ai

a hospital bill auditing app built at hacktj, where it won best mobile hack. it accepts photos and pdfs, extracts medical codes and prices, checks them against cms data using rag, flags suspicious charges, and drafts a dispute letter.

pokepy

a small language model written from scratch in numpy to generate pokémon names. i implemented the forward and backward passes without ml frameworks to learn about weight initialization, activation stability, and gradient flow.

awards

  1. best use of exa · $1,000

    zero human company hackathon nyc · issued by terac

    built orit, a fully autonomous company that simulates product performance in the market.

  2. best use of gcp · $50

    tillyhacks · issued by tillyhacks

    built felix, a multi-agent system for real-world computer actions.

  3. best mobile hack · ~$120

    hacktj · issued by hacktj

    built fern.ai, an ai hospital bill auditor.

future

i want to keep learning by:

  • studying how models connect perception, memory, and action
  • understanding how learned representations generalize to new situations
  • reading existing work and testing ideas through small projects