FOUNDER
Who is Ugyen Dendup?

Most teams building AI agents today are stuck at the same point: great demos, broken production.

Ugyen Dendup is building Hopeless, a control layer that sits between AI agents and your backend so agents can safely move from pilot to production—without taking your infra or costs down in the process.

From Hydropower Sites to Bhutan’s First AI Startup

Ugyen is 25 and in his final year studying computer science (AI & data science) in Bhutan. A little about Ugyen:

• Grew up near hydropower construction sites where his parents worked

• Won a scholarship to study in the capital and went deep into STEM

• Founded Bhutan’s first student‑run robotics club, a nonprofit to give other students access to robotics and STEM opportunities

In college, he met his co‑founder Jamphel as a dormmate. They started hacking on projects together and, in 2022, built an ML chatbot before the current LLM/agent wave.

They entered a national hackathon in Bhutan and placed as runners‑up, winning a small USD grant.

Instead of treating it as prize money, they treated it as seed capital and started NoMindBhutan, which they believe was Bhutan’s first AI startup.

COMPANY
NoMindBhutan…then Hopeless

NoMindBhutan transformed from a project into a real business by:

• Building internal AI tools and agent‑like chatbots

• Working with top banks, airlines, and government agencies in Bhutan

• Operating as a B2B services company, customizing AI solutions

They even got featured on Bloomberg, which gave them early credibility.

But structurally, it was still a services business:

• Hard to scale without adding more people

• Not a clear venture‑scale, product‑first opportunity

They kept serving clients but looked outward. That led them to Entrepreneurs First (EF) in Bangalore, where they joined the Fall 2025 cohort and started building their second company: Hopeless.

The Problem: Agents Think in Goals, Backends Think in Requests

Ugyen’s core observation:

Why?

Most backends today were designed for human operators, not autonomous agents.

• Humans interact request by request

• Agents operate goal by goal and generate bursts of parallel requests

Example:

• A human checking a payment:

  • Open one user → open payments → 1–2 backend calls

• An AI agent checking payments:

  • Get a goal like “check the payment for that user”

  • Fetch many users → then fan out 100–200 requests to check transactions

This creates:

• Fan‑out traffic patterns

• Rate‑limit issues

• Database overload

• Unpredictable and high LLM/API bills

Right now, most AI agents are still demos or internal pilots, so teams haven’t fully felt this pain. Ugyen’s bet is that as agents move into production over the next 1–2 years, infra failures and cost explosions will become a first‑order problem.

The Solution: Hopeless as the AI Control Layer

Hopeless is a smart API middleware between AI agents and your backend.

Think of it as the control, safety, and optimization layer for agent workloads.

Early focus areas:

• Token optimization

  • Reduce redundant or inefficient LLM usage

  • Keep agent behavior within cost budgets

• Fan‑out control

  • Manage and throttle the “hundreds of parallel requests” agents naturally create

  • Protect databases and internal APIs from burst traffic

• Secure parallel execution

  • Coordinate multiple agent calls safely

  • Respect rate limits and backend constraints without manual patching

The goal:

Who’s Using Hopeless Today

Hopeless is already live with:

• 113+ users

• 10 paying customers

• 13 active users building and running agents through the platform

Two main user profiles (ICPs):

1. Solo AI developers & freelancers

   • Building agents for clients or products

   • Need guardrails so infra doesn’t fail and costs don’t spike as usage grows

2. Internal teams at fintechs and enterprises

   • IT / platform / data teams building internal AI agents

   • Those wanting to connect agents to production systems safely, without risking outages

In both cases, Hopeless becomes the intake and control layer for agent traffic into existing infrastructure.

Where Hopeless Is Going

Near‑term vision:

Long‑term:

As the market for AI agents grows, infra teams will need dedicated safety and control layers for agents, the way they needed load balancers, CDNs, and observability in previous waves

Hopeless is positioning itself as that layer.

If you’re building agents that touch real systems—and you’re worried about stability, rate limits, or cost blow‑ups—Ugyen is building Hopeless for you.

TL;DR

Ugyen Dendup is a 25-year-old Bhutanese founder and EF Bangalore fellow who went from building the country’s first AI startup and robotics club to creating Hopeless, a control layer that makes backends AI‑ready so AI agents can safely move from pilot to production.

Follow Ugyen
I want to be featured on foundersbrief
Follow foundersbrief