Hello : )
I’m Kiran Jesudasan, Head of Design at Conode, and this is my unpaid intern, Meesh. Here are my ramblings.
I’m a designer, and designing for ambiguity, transparency, and validating AI is what drives me.
Says every other designer in the world, right? Obviously. Here’s what makes me different.
The hardest (most fun) problem with designing AI isn’t presenting answers.
It’s knowing they’re true, proving it fast, and communicating that truth in a way people can actually trust.
I joined as employee #1, tasked with giving a deeply technical knowledge-graph platform a soul. I am Head of Design and work with a team of 4 to define UX, UI, research iniatives and brand strategy.
Seven years later—after real-world deployments and an acquisition—I’ve learned a bit about the fascinating intersection of Design and grounding AI.
In 2018, we had no users, no customers, and no VC funding.
By 2026, we secured 2 rounds of VC funding, won several rounds of grants, and global customer traction.
What began as software for autonomous-vehicle testing evolved into a platform used across logistics, insurance, and SMBs—enabling teams to deploy domain-specific LLMs grounded in their own knowledge graphs.
We’ve successfully exited. I’m now looking for the next challenge.
Ok but what’s the impact?
Designing AI: Eight Commandments/
Habits from the Mountain.
The missing two were casualties of iteration.
Design is a flexible process, not a panic attack. It can scaled up or down, sure, in fact one could argue that you must, but it’s not “ship a hi-fi prototype by tomorrow, from nothing.” Skip entire phases and it’ll bite you. (Like my intern does).
Research
Thou shalt not cosplay as The Answer Guy. Talk to customers early. Ask dumb questions. Be vulnerably curious. The sooner you learn, the sooner you stop building the wrong thing with confidence.
Know thy user (for real). Not what they say in the 30-minute intro call—what they do today. Observe. Follow. Map out. Remove bias. Find the patterns they don’t know to reveal and design for that. Validate the experience; this will shield you and your work from the 'executive who read a TechCrunch article' syndrome.
Ideation
Use AI. Then use your brain. AI can speed you up, but it can also miss the actual theme, especially during research. Always do a human pass—because missed themes = missed innovation.
Ideate like you mean it. AI is decent at remixing, not originating. Your job is volume: 50–100 messy, based on , napkin-grade options before you commit to a “concept.”
Validation
Stay low-fi longer than feels comfortable. Post-its → chicken scratch → coherent concepts. Then let AI help you move to mid-fi. Not the other way around.
Validate with the right fidelity. Hi-fi isn’t “more correct.” Sometimes unbranded mid-fi is better: less noise, clearer signal, faster learning.
And ya know... the rest
In 0→1 organizations, design is a contact sport. Be bold. Define the ideal flow before it’s engineered—then collaborate and map out constraints. Design roadmaps of how to get to the ideal, not just pretty prototypes of it.
Designed India’s first ultra-cheap prosthetic hand.
Imagined how Uber’s flying car program would work and what the experience is like.
Designed underbody parts for Volkswagen India’s lineup, which saved 3000 cars/year needing their engine replaced.
Been on the jury of the 2025 AI Design Award
Some other neat
things I've done
Life stuff...
I've lived in 4 countries for at least 3 years each, India, the US, Bhutan and Malaysia.
I play the drums in a metal band
I'm big into running, cooking, and testing spice tolerances.
I nerd out on DnD, Warhammer 40k and most strategy games
I'm slowly chipping away at my private pilot's license.
Copyright Kiran Jesudasan 2026