Dima Perkis — dimadimadima
DIMAPERKIS
Atlanta dad and recovering strategy consultant with newfound AI superpowers
I help people discover what AI can do for them and put it to work …
- reducing daily friction
- freeing up time
- growing their business
Build something with me ↗ See what I’ve built ↓
Recent Projects & Experiments
147,594 Spotify streams
178 Live shows reconstructed
5 Taste dimensions
112 ATL events in next 90 days · Aug 21](/content/atl/index.html)
I made an event radar to find music and live events I would love
Radar started as a way to find the kinds of events that actually bring me joy and make life in Atlanta more fun. I pulled together fourteen years of Spotify history, a reconstructed concert log, venue preferences, family constraints, and the patterns behind what reliably feels worth leaving the house for.
The other half is relational. Once the scoring exists, it becomes easy to send someone a live link and say: this feels like us, want to do it?
It helps me live a more fun life and stay closer to people I want to see more often.
What is there What could be
\
Can AI make a high-value service feel personal before the first call?
I’m exploring that question through dimaperkis.com with high-end home service businesses.
One example starts with an ordinary property photo. AI turns it into a visual concept, personalized outreach, and a proposal the homeowner can explore.
The homeowner sees what could change before the first sales call ever happens.
Live motion preview
15 games](https://wiggleworks.vercel.app/)
A webcam became a playroom
My kids needed more ways to move. Wiggle Works turns their bodies into the controller for balloon popping, freeze dance, ninja slicing, magic paint, and eleven other games.
No controller. No backend. No need to sit still.
2 kids 15 games 100% on-device
What do 222 hours of my own voice know that I don’t?
I pulled together meetings, coaching sessions, dictations, and voice memos—then transcribed, classified, and made them searchable. The system can show how my language changes by setting, who brings out different parts of me, and which commitments keep resurfacing.
It has already shown me that I hedge almost twice as much in meetings as I do when I think out loud alone.
[Voice Intelligence indexed
_SW_ Superwhisper_G_ Google Meet_Z_ Zoom_DG_ Deepgram Nova
1,438 recordings
1.95M words
8,688 searchable chunks
One finding **I hedge almost twice as much in meetings as I do when I’m thinking out loud alone**](/content/projects/voice-intelligence/index.html)
Useful, delightful, or both
See the full project archive →
Road-trip stops that work for toddlers \
1,855 places to move, eat, and reset between Atlanta and Charlotte—organized around the trip we actually take
Open project ↗](/content/parks/index.html)
The power bill was blaming the wrong hours \
10,487 hourly readings showed the real problem was HVAC and the dryer colliding on hot weekends
Open project ↗](/content/projects/power-analysis/index.html)
A reading list that knows what is on the shelf \
Dewey checks every recommendation against our actual Fulton County Library branch before we leave home
Open project ↗](/content/projects/dewey/index.html)
The dashboard behind the dashboards \
Brain Ops watches 18 automated jobs and tells me whether they are merely alive or actually doing their jobs
Open project ↗](/content/projects/brain-ops/index.html)
Which properties look like real opportunities? \
Permit Propensity combines 8,701 permit records, 143,031 parcels, and valuation data into a more useful view of who to talk to next
Open project ↗](/content/projects/permit-propensity/index.html)
What 14 years of listening says about me \
147,594 Spotify streams became a taste model, a concert history, and the engine behind On Dima’s Radar
Open project ↗](/content/projects/spotify-analysis/index.html)
What would you point AI at first?
That’s where we start. We find the version worth building, and you come away with something you can actually try.