JINGHAN MA · FOUNDER, WHAISPER
A self-taught AI engineer’s journey from Wall Street spreadsheets to building a delegation system that gives everyone an agent of their own.
Start at the beginning
NOT GIVEN
My parents ran a small business and were always working. The ride home from school was one of the few times I had my mother to myself, so from the back of her bicycle I asked everything: why the moon only came out at night, why windmills turned, and why she and my father were always too busy to spend more time with me.
Curiosity began in the small spaces between their workdays.
NOT FOUND
I arrived in New York for a finance degree, then entered Wall Street. From the 30th floor, I could see the Statue of Liberty. Inside, repetitive derivatives work consumed every hour for reflection. I turned down Morgan Stanley's full-time offer and sponsorship. Then I left Goldman Sachs too.
At night, fireflies in Central Park reminded me that I wanted to make my own light.
SPARK
A Google VP spoke about AGI. For the first time in years, the future felt larger than the path in front of me. I began teaching myself Python, machine learning, and software engineering. Workdays were for finance. Nights were for code.
The light was not found. It was practiced, one night at a time.
REBUILT
I moved west, rented an $800 room, audited a machine-learning bootcamp, and learned at night. By day I turned contract work into production systems: pricing models for global airlines, AI pipelines for Panasonic Avionics, then a data team built from zero at Helium 10.
I treated every job as a paid learning opportunity.
SCALE
I joined Siri AIML and later built agentic evaluation systems for end-to-end task quality. Working beside world-class engineers taught me rigor and scale. It also clarified my founder thesis: assistants improve when we measure real human outcomes, not vanity metrics.
From turn-level accuracy → to task-level completion
PROOF
At the UC LAUNCH Accelerator, the personal-agent prototype won the Audience Choice Award and earned backing from The Sega Sammy Fund. An earlier version had already placed third at a hackathon. Each test made the same signal louder: people do not need another chat box. They need an agent that can carry work across the finish line.
GROWN
At Apple AIML, I learned what it takes to evaluate intelligence across billions of devices, and why assistants still fail at the outcome. In May 2026, I started Whaisper, an AI delegation system that gives every user a personal agent with its own phone number. You hand it a task by voice or text. It works across the digital and physical world, follows through, and returns only when the task is complete or the outcome is clear.
Said, done.
THE FOUNDER THESIS
I left prestigious jobs because repetitive coordination was crowding out the work only humans can do.
At Apple, I learned to evaluate complete tasks and find the long-tail gaps hidden by aggregate metrics.
Almost every clinic, vendor, airline, and landlord has a phone number. A personal agent should be able to call, follow up, and finish the work.
THE NEXT CHAPTER
Whaisper gives every user a personal AI agent with its own phone number. Delegate by voice or text, then stop managing the task. The agent works across email, calendar, messaging, the web, and phone calls until there is a verified outcome. Said, done.