# Utkarsh Priyadarshi

> AI Builder at Snowflake · Menlo Park, CA

Official portfolio: https://priyadarshiutkarsh.com/about/

## About

I’m an AI Enterprise Builder at Snowflake with a part-time focus on cybersecurity. I build AI agents that solve real business problems and eliminate repetitive work, for any function across the company.

I’ve gone deep into marketing and branding too, because a great product and the story around it are two halves of the same craft. I believe in a conqueror mindset: if you can dream it, you can achieve it.

Outside of work, I quietly support causes I care about: education, elder care, and sustainability.

## Experience

### AI Builder (Office of the CIO) — Snowflake

Sept 2026 - Present

Learning Snowflake & delivering agentic solutions: Enterprise Technology team.

### Scientific Researcher — AI Safety Camp

Jan 2025 - Apr 2025

Worked on activation maximization and collaborated with sub team of sparse autoencoders.

### Undergraduate Research Assistant — UW Madison Computer Sciences

Sep 2024 - Dec 2024

Explored CDVAE model for predicting materials, fine tuning, and structural graph impacts.

## Projects

### RageBaitify

The AI defends a terrible take. You argue it down.

An argue to learn debate site. The AI defends a deliberately provocative take and you try to break its conviction. Only solid, evidence based points move it, and every claim is backed by real Wikipedia sources.

Technology: AI, Debate, Wikipedia

Project: https://ragebaitify.com

### Corenous

Your Mac never forgets. And it never tells.

A local memory engine for macOS that quietly remembers everything you see, copy, and read. Ask in plain language and it finds the exact moment you need, all on your own machine with no cloud.

Technology: Python, Apple MLX, Qwen2.5-VL, SQLite

Project: https://github.com/priyadarshiutkarsh/corenous

## Research

### Mechanistic Interpretability Via Learning Differential Equations

Blogpost on interpreting transformer representations in time-series prediction.

Source: https://www.lesswrong.com/posts/qdxNsbY5kYNqcgzFb/mechanistic-interpretability-via-learning-differential

## Education

### Bachelor of Science in Computer Science — UW Madison

2023 - 2026

Minor in Educational Policy Studies & Social Justice.

## Skills

- Agentic AI Systems: Multi-Agent Systems, Agent Orchestration, Tool Use & Function Calling, RAG, Prompt Engineering, MCP, Workflow Automation
- Cybersecurity: Secure AI Systems, Threat Modeling, LLM Security, Application Security
- AI-Assisted Development: Coco, Claude Code, Codex, Cursor, Windsurf, Lovable, Supabase, Vercel
- Product, Design & Branding: Figma, Google Stitch, Product Design, Branding, Marketing, Storytelling

## Contact and profiles

- Email: priyadarshiutkarshofficial@gmail.com
- GitHub: https://github.com/utkarshapex
- LinkedIn: https://www.linkedin.com/in/utkarshprriyadarshi/