How a Private Chef Startup Went All In on AI Agents

Yhangry is a private chef marketplace doing $15 million in GMV that's going all in on AI agents across every function of the company — from an autonomous bug fixer that shipped 25 fixes in its first week to an AI product that matches chefs and customers instantly. In this recent batch talk, founder Siddhi Mittal walks through three real business use cases for agents at yhangry, how she turned teaching AI in plain English into a growth channel worth $50K in free conference slots, and the hard org decisions she made to rebuild the company as AI native from the ground up.

Siddhi Mittalguest

May 19, 20264mWatch on YouTube ↗

WHAT IT’S REALLY ABOUT

Private chef marketplace accelerates growth by embedding AI agents everywhere

Yhangry built an autonomous bug-fixing agent in under four days to clear a long-tail backlog of low-priority bugs while aiming to 10× growth from roughly $15M GMV.
The founder turned “teaching AI agents in plain English” into a marketing wedge, securing about $50K of conference speaking slots for free and funneling that attention back to Yhangry.
The company is developing “Yhangry AI” to collapse multi-step chef booking workflows into faster, higher-confidence matches using historical customer/chef interaction data.
On the chef side, an assistant vision (“Claude for chefs”) aims to eliminate repetitive admin and messaging, but variability in chef needs is pushing the team to find a workable MVP.
Operationally, Yhangry is reorganizing to be AI-native—running weekly agentic labs, upgrading engineering leadership, and standardizing how teams build and evaluate agents.

IDEAS WORTH REMEMBERING

5 ideas

Use agents to tackle the “dropped” backlog work humans avoid.

Yhangry targeted small, annoying bugs that never reached the top of the ROI stack, shipping 25+ fixes in week one to improve product quality without diverting engineers from higher-leverage initiatives.

Context is the core constraint for autonomous engineering agents.

The team’s key learning is that pass rates depend on feeding sufficient codebase and product context so the system can improve rather than repeatedly fail on edge cases.

Teaching can be a scalable acquisition channel when paired with a subtle product funnel.

By offering a 30-minute “how to build AI agents” talk, the founder earned free conference slots, then embedded affiliate integration and a Yhangry AI demo inside the deck to convert attention into demand.

AI can compress multi-step marketplaces by replacing back-and-forth with confident one-shot matching.

Yhangry sees booking friction (messages, delays, repeated info) as a major conversion leak and believes its historical data can enable faster, higher-quality chef recommendations.

Chef-facing automation must handle highly divergent workflows to succeed.

Some chefs want structured, “by-the-book” flows while others ask for open-ended help, making MVP scope selection and guardrails critical before shipping broadly.

WORDS WORTH SAVING

5 quotes

So built autonomous bug fixer in less than four days, literally 25 plus bugs fixed in week one and shipped.

— Siddhi Mittal

So I got $50K worth of conference slots for free.

— Siddhi Mittal

There is no reason for this back and forth to take days.

— Siddhi Mittal

It's like Claude for chefs.

— Siddhi Mittal

I fired my tech lead 'cause I realized he did not know what skills was, and he was the ceiling in our company.

— Siddhi Mittal

Autonomous bug-fixing agentBenchmarking one-shot bug fixes (60–70%)Founder-led education as growth channelConference talks as customer acquisition funnelAI-driven chef–customer matchingChef-side admin automation (“Claude for chefs”)AI-native org design and weekly agentic labs

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.