Jeremy Bischoff

I turn expert workflows into production AI agents. Founding engineer at WarmApp, an AI coaching mentor for gymnastics and calisthenics. Former NCAA gymnastics champion and USA national team member.

Try WarmApp Work with me

About

WarmApp

An AI coaching mentor for calisthenics and gymnastics athletes. It programs personalized workouts, guides each session, remembers your training, and adapts the next session based on what came before. Built on the methodology of my Olympian co-founder Ruben Lopez: the same system I trained under at Stanford and the one behind various world champions and Olympians. We've spent the last year codifying it into software.

Try it at warmapp.ai

Background

Stanford University: B.S. in Computer Science (AI Track), Minor in Management Science & Engineering
Elite Gymnast: 4x NCAA Champion, All-American, USA National Team Member, Olympic Trials Qualifier

Current Position

Software Engineer, Watters & Jacobs, LLP / Raellic R&D

Co-designed and built MOSAIC, a covert-channel system that fragments files across email headers and includes companion detection heuristics. Co-authored the white paper; demonstrated to the FBI.
Designed and built an internal, secure firm operations platform — email, calendar, billing, time tracking, financials, document handling, and case management in one in-house system.
Named co-inventor on a utility patent pending. Filed with the USPTO in 2026.

Experience

AI Tools & Utilities Engineer, Seagull Work / Imantics, Inc.

Built business development and schematic design agents for an electronics design AI company and an IoT company.

Software Development Intern, Seagull Concepts, LLC

Software development for electronic schematic visualization.

Download resume

How I Build

I turn expert workflows into production AI agents. Each one trained to the standard I learned in elite gymnastics: every tenth matters.

Honest. The agent admits what it doesn't know.
Anchored. It fits the expert's existing workflow.
Minimal. Every surface earns its place.
Refined. The edge cases are the bar.
Durable. Learns from every case it sees.

See the method