Rent
Hourly or daily at malls, events, and destination retail.
Entry wedge
A personal follow-me rover
You walk.
It follows.
Your hands are free.
Built for shopping first.
Designed to carry wherever people do.
Vedant Parnaik
Robotics AI Engineer · Los Angeles
01 / The problem
Shopping ends when your hands are full.
You stop buying. You make two trips to the car. You skip the mall because hauling bags hurts.
For older adults, people with back pain, or parents with children, carrying is not an inconvenience. It is the limit.
“The cart stays in the store.
The bags come with you.”
02 / The insight
People think this is a
self-driving problem.
It isn’t.
YOU
are the navigation.
A follower needs no maps, route planning, per-store setup, or building integration. You walk; it stays with you. That reframing removes the deployment problem that kills indoor robotics.
03 / The product
A rack that
stays with you.
- 01No store app or integration
- 02Works across shops and to the car
- 03Stops when identity is uncertain
- 04Designed around commodity hardware
04 / Working demo
Not a render.
An early rig,
already moving.
Filmed live in a yard: the rover follows me with no scripted path. Vision runs on a laptop over Wi-Fi, and the rover stops its motors if commands stop for 0.25 s. Straight-line following holds up; sharp side cuts can still lose me.
Also working, in simulation so far: re-identification and obstacle avoidance. Not built yet: the load-bearing bag rack and onboard compute.
05 / What exists today
Early, working,
and honest.
This is a prototype, not a product.
- Drives4-motor mecanum base
- SeesSub-second live camera stream
- RespondsKeyboard or game controller teleop
- Fails safeMotors stop if commands stop for 0.25 s
- FollowsLocks onto one person and follows at about 2.6 m
- Re-identifiesCatches tracker ID swaps and finds the right person again (in simulation)
- AvoidsUltrasonic and camera obstacle avoidance (in simulation)
06 / What is actually hard
The product is trust.
The failure mode that matters is not “it got lost.” It is “it followed a stranger.”
In a crowded aisle, it must stay locked through occlusion—and refuse to move when it is not sure.
Stop on uncertainty
07 / Why this is not a smart cart
Smart carts
Belong to
Belong to
the store.
- Chain-specific checkout
- Stays inside one location
- Capex and rollout per store
This rover
Belongs to
Belongs to
the person.
- Moves between shops
- Continues to the parking lot
- Ready to use anywhere you shop
Different buyer. Different product. A much lower barrier to the first customer.
08 / Business model
Rent first.
Learn fast.
Own
For people who haul regularly—or cannot haul at all.
Consumer scaleBuild on it
Sell the follow-me base as a platform for developers. The brain is already open source as followme-bot.
Platform upsideRental keeps the price barrier near zero and puts units in front of real users quickly.
09 / Expansion
One capability.
Many loads.
01Older adults
Independent shopping
02Hospitals
Supplies down corridors
03Technicians
Tools across factory floors
04Travel + events
Airports and markets
This category should look like the mobility scooter, not the warehouse robot: affordable enough to own, trusted enough to walk behind you.
10 / Roadmap
From target lock
to trusted companion.
Three engineering gates to a real-world pilot.
Prove the loop
- Teleop + target lock
- 4-motor base
- Bench and hallway testing
Follow a person
- Person-following on the rover
- Identity re-acquisition (working in simulation)
- Load-bearing rack
- Onboard compute
Earn trust
- Obstacle safety (working in simulation)
- Stop on uncertainty (working in simulation)
- Pilot rentals in real aisles
11 / Why me
I build through
the messy middle.
I built Handme, a mobile manipulator whose base, arm, and hand had to work together in real time. Every delay and calibration error exposed what the plan missed.
That was version one of this journey. I am a robotics AI engineer with six years across perception, sensor fusion, SLAM, and embedded systems—and a room full of electronics that rarely behave the first time.
View engineering portfolio ↗12 / The ask
Help me move it
from the hallway
to the aisle.
Raising a first check to build the load-bearing chassis, move vision onboard, make person-following reliable, and put prototypes in front of real shoppers.