Waave
US taxi platform with an NTEP-certified meter app.
CTO · 2019–2025 (NYC/Boston)
Read the full case study
The problem
A taxi meter running on a phone is a legal measuring instrument in the US. Without NTEP certification the app cannot legally charge a fare, and investors doing due diligence want to see both the certificate and a cloud bill that makes sense.
What we built
Led the taxi meter app through National Type Evaluation Program certification with measurement error under 0.1%. Ran a full AWS inventory and cost audit, then a migration to a new AWS account with a run-rate around $1,185/month after cleanup. Rebuilt the platform for investor due diligence and set the AI roadmap: demand prediction, driver routing, dispatch agents, AI QA for the mobile apps, fleet rebalancing.
Decisions that mattered
Certify the app, not a hardware meter
Hardware meters certify easily and cost per vehicle. Getting software certified was slower and harder to prove, and it removed a per-car hardware line from the unit economics permanently.
New AWS account instead of cleaning the old one
Years of untagged resources made the old account impossible to reason about. A clean account with infra-as-code cost a migration but gave a bill that could be explained line by line in a data room.
AI roadmap sequenced by data readiness
Dispatch agents were the exciting item. Demand prediction went first because the ride history to train on already existed and the routing work depended on it.
Numbers
<0.1%
meter measurement error
NTEP
certified in the US
~$1,185/mo
AWS run-rate after cleanup
Stack
Related work
Got a product that needs to go further?
Fixed-price scope in the first call. No discovery phase invoice.