click to light
[ 01 ]Software, built with care

I build
like it matters.

I’m Aakash Dhanyamraju — a software engineer who treats every product like it’s the only one. I obsess over the edges, ship with intent, and stay until it’s right.

Scroll
Ship outcomes, not commits
Care is a feature
Quiet hands. Loud impact.
If it’s worth doing, do it twice
Own the bug. Own the fix.
Build like someone is using it tomorrow
Ship outcomes, not commits
Care is a feature
Quiet hands. Loud impact.
If it’s worth doing, do it twice
Own the bug. Own the fix.
Build like someone is using it tomorrow
Ship outcomes, not commits
Care is a feature
Quiet hands. Loud impact.
If it’s worth doing, do it twice
Own the bug. Own the fix.
Build like someone is using it tomorrow
[ 02 ]Manifesto

Idon'tshipcode.Ishipoutcomes.Itreateverylinelikeit'llbereadat3a.m.bysomeonetiredandtryingtofixsomethingImissed.Iwriteless.Ichecktwice.Istayuntilitworks,andthenIstayalittlelongertomakesureitstaysthatway.

A working philosophy, not a slogan.
[ 03 ]Impact

The work, in numbers that survived production.

01
0.0%

Uptime maintained across production deployments

02
0K+

People served by systems I’ve built and kept running

03
0%

Latency cut on the workflows that mattered most

04
0M+

Database records orchestrated, day in and day out

Numbers don’t make a craftsman, but craftsmen tend to leave numbers behind. These are mine, earned across products real people use.

[ 04 ]Selected work

Stories, not stacks.

What I built, why it mattered, and what was different the day after I shipped it. The tools are on the resume.

012024 – Now

An autonomous agent platform that doesn't lose its mind in production.

A multi-tenant control plane for deploying AI agents that reason, act, and stay observable. Built so a non-engineer can launch one in minutes, and an engineer can debug one at 2 a.m. without crying.

My role

Architecture, end-to-end build, the late-night fixes nobody sees.

Outcomes
  • Scaled to 50K+ daily interactions without paging anyone
  • Cut end-to-end latency by 60% on the hot paths
  • 99.9% uptime, and the boring kind of operations
022024

Pick-A-Pad AI — turning a slow, painful review into a 60% faster decision.

A tenant screening platform managing thousands of property units. The bet: let an LLM read the messy documents so humans can make the actual call. The hard part was making something operators trust enough to use every day.

My role

From a blank Figma to a system property managers depend on.

Outcomes
  • Managing 2.9K+ property units in active use
  • Cut manual document review by ~60%
  • Quietly removed the part of the job nobody liked
032023

A social platform that grew to 500K users — without ever feeling slow.

Joined as an intern. Left having co-built the recommendation engine, the moderation system, and most of the surface area people actually touched. The win wasn't being clever; it was making a fast app stay fast.

My role

Owned the hard performance work, mentored two engineers along the way.

Outcomes
  • Supported 500K+ users, 2M+ records, zero downtime weeks
  • Cut API response times by 35%
  • Engagement up 25% after the recommendation rewrite
[ 05 ]Principles

How I show up every day.

Not values on a wall. The actual operating system behind the work.

i.

Ownership without a ticket.

If I see it, it's mine. The bug, the doc gap, the awkward UX. Waiting for someone to assign it is how things rot.

ii.

Boring on the outside.

Software should feel like a good appliance. Predictable, quiet, durable. The cleverness lives inside, where it earns its keep.

iii.

Read it twice. Ship once.

Reviews aren’t a formality. They are the last place the future-me will thank the present-me. I show up to them.

iv.

Speed is a side-effect of clarity.

I don’t go fast by skipping steps. I go fast by knowing exactly which steps I’m skipping, and which ones I’m not.

v.

Learn out loud.

If I figured something out, somebody on the team should be able to figure it out in half the time tomorrow. Notes, PRs, docs, voice.

vi.

Care is a feature.

Loading states. Empty states. Error copy that doesn’t condescend. The 1% that decides whether people stay.

[ 06 ]The way I work

A repeatable five-move game.

1

I read first. Listen second. Type third.

Before any code, I want to know what hurts, who it hurts, and what 'done' actually feels like. That conversation is the project.

2

I draw the smallest thing that could be wrong.

A diagram on paper, a Loom, a doc. Fast and disposable. I'd rather find the flaw in 20 minutes than three sprints.

3

I ship something ugly within a week.

Real shape only emerges when something actually runs. Polish lies. Behavior tells the truth.

4

I sit with the people using it.

I watch where they pause. Where they squint. Where they alt-tab. That's where the next version begins.

5

I clean up after myself, always.

Dashboards, alerts, runbooks, deprecations. I treat the project as alive after launch — because it is.

[ 07 ]A short letter

Hi,

I’m writing this because a resume can tell you what I’ve built, but not why I’ll be a good person to build with.

The thing I’m most proud of isn’t a project. It’s a habit: I don’t leave things half-done. Not the PRs, not the alerts, not the documentation, not the awkward conversations.

Nearly three years in, I still approach every standup with the curiosity and energy of day one—because if the project doesn’t feel interesting, what’s the point?

If you’re looking for someone who will quietly raise the floor of your team — who will write the doc nobody asked for and own the bug nobody saw coming — we should talk.

— Aakash

[ 08 ]Let's build

Have a problem
worth solving?

I’m open to full-time roles where craft is taken seriously, and to occasional project work that’s a real challenge. If that’s you, write a line below — or just email me directly.

GitHubLinkedIn
LocationRemote · India
Reply timeUsually same day
StatusOpen to new work

Hitting send opens a prefilled email in your own mail app — just review and send. I read every one myself.