Arjun Kannan

Co-founder at ResiDesk

Arjun Kannan

I build products for messy systems.

I co-founded ResiDesk. We build AI tools for multifamily operators and use resident conversations to help teams see problems earlier, respond with more context, and make better decisions.

Before that I worked at Climb Credit and BlackRock. Before that I studied applied physics at Cornell and got into software in a lab because writing code turned out to be the fastest way to make a tedious process better.

From the outside those jumps can look random. They are not. I keep ending up in the same kind of problem: the information is already there, but it is trapped in the wrong system, or it reaches the wrong person too late.

ResiDesk Writing Selected links
Arjun Kannan
A good place to start Spring 2026

If you are new here

Now

I run product and engineering at ResiDesk.

Before

I built products at Climb Credit and BlackRock, then brought the same instincts into housing.

Start here

If you want the shortest version, go to Work, then Recent, then Conversations.

What I care about

Products that help someone make a better decision in a messy real workflow, not a clean demo.

About

How I got into this

Harnoor interview TechTimes

My original plan was academic. Both of my parents have PhDs, and for a long time I assumed I would do the same. At Cornell I studied applied physics because it sat in the middle of the things I liked most: fundamental science, real systems, and hard problems that still had some contact with the world.

Software showed up sideways. I was working in an electron microscopy lab and wrote code to make magnetic-noise setup faster. It cut down manual work, it was useful right away, and it was the first time I felt how powerful software is when it closes a loop for someone. That pulled me out of the original plan.

The industries have changed since then. At BlackRock I worked on making institutional systems more usable for advisors. At Climb it was underwriting against outcomes instead of surface-level safety. Now at ResiDesk it is turning everyday resident conversations into something operators can act on before issues get expensive.

Cornell

Software was never the plan.

I was working in an electron microscopy lab and wrote a small tool to speed up magnetic-noise setup. That was the first time software felt obviously useful to me.

BlackRock

Context mattered more than the puzzle.

I came back six months later, interviewed again, and ended up in New York. It was also the first place where I saw how much better I work when the problem has real context.

Climb Credit

The question we were asking was wrong.

Instead of asking who looked safest on paper, we asked what happened to earnings after the program. That changed underwriting, product, and data, and helped annual loan volume grow from $1 million to $300 million.

ResiDesk

Start with the resident.

Residents already tell operators what matters. The job is to make those conversations useful earlier, for the resident and for the operator, before the issue gets expensive.

Work

What I have built

ResiDesk TechCrunch Law360

Most of the work has been some version of the same job: take a messy system, figure out where the missing context is, and turn it into something a person can act on.

ResiDesk

7%

Higher renewals and rent after getting closer to what residents were saying and acting on it earlier.

Climb Credit

$1M → $300M

Annual loan growth tied to underwriting around outcomes instead of a narrower picture of risk.

BlackRock

$40M ARR

Retail analytics product from zero to $40 million in ARR in year one.

ResiDesk

Co-founder

We help multifamily operators understand what residents are saying across renewals, rent, maintenance, and staff issues. The product matters because it helps teams act earlier and with more context, not because it looks clever in a demo.

Climb Credit

CTO and CPO

We built around a different underwriting question: not who looks safest on paper, but what happens to earnings after the program. That changed who got access, what data mattered, and how the product had to work.

BlackRock

Product and engineering

The work was turning institutional infrastructure into products advisors could use with clients. Same underlying information, better interface, better decision.

Recent

Recent writing and press

HackerNoon TechBullion TechTimes
TechBullion article preview

TechBullion / May 6, 2025

NOI, Not Noise

A clean summary of how I think about AI inside operations, resident signal, and the difference between noise and something a team can act on.

Law360 / January 8, 2025

ResiDesk Simplifies Resident Engagement For Deeper Results on signal, retention, and operations.

20for20 / November 20, 2025

White paper on what multifamily residents think about Wi-Fi and what operators can learn from the sentiment underneath it.

HackerNoon / September 18, 2024

The Power of Applied AI on practical AI inside real estate workflows.

TechTimes / August 20, 2024

Feature on the arc across real estate, finance, and education.

TechCrunch / April 13, 2016

Coverage of Climb Credit’s outcome-based underwriting model.

Conversations

Interviews and talks

Harnoor Greg Kamradt 20for20 BuiltWorlds

If you want the less edited version, start here. These are a better way to hear how I think than any summary I could write about myself.

20for20 / February 2026

AI sentiment analysis: how residents really feel about Wi-Fi

The talk version of the resident sentiment work.

BuiltWorlds / March 2026

Investing from the owner and operator perspective

A room full of owners, operators, and investors thinking about what AI looks like in practice.

TechBullion / May 2025

NOI, Not Noise

A good summary of what resident signal means when it is tied to operations.

Writing

Writing

Substack buy the car, not the engine Phones and Headphones

On Substack I usually write when the obvious framing feels wrong. Lately that means AI, product loops, advice, and the gap between a good demo and a system that survives contact with real work.

Read the archive

The bar is simple.

If a tool cannot help someone move faster toward a better decision, it is probably not doing enough.

Start with the workflow.

If you do not know what someone is trying to get done, you are mostly decorating the surface.

Buy the car, not the engine.

The model matters less than the system around it: context, tools, guardrails, and where it sits in the workflow.

A demo is not a product.

The interesting question is whether people keep using it in a messy real workflow, not whether it looks impressive for three minutes.

FAQ

A few things people usually ask

What kind of AI work do I do?

I build AI products that sit inside real workflows. At ResiDesk that means resident sentiment analysis, AI copilots for property teams, and agent systems that help operators act earlier with more context.

What have I built before ResiDesk?

Before ResiDesk I worked at Climb Credit and BlackRock. At Climb, annual loan volume grew from $1 million to $300 million. At BlackRock, a retail analytics product went from $0 to $40 million ARR in its first year.

What do I usually write and speak about?

Mostly AI products, agent workflows, LLM evaluation, product loops, and the difference between a good demo and a system that survives contact with real work. The recent 20for20 talk used resident conversations across 30,000+ units in 11 states as the starting point.

What is my view on AI?

I care more about whether a system helps someone make a better decision than whether it looks magical for five minutes. That usually means context, evaluation, and human-in-the-loop design before full automation.

Investing

Investing and mentoring

Advice is a System, Not a Person

I have invested in more than 100 startups and mentored through Techstars. I am usually drawn to founders who are close to a real problem, close to the customer, and honest about what is still unclear.

That matters even more now. Generic advice is cheap. What is still useful is context, judgment, and helping someone get to the next real decision faster.