Turn “hunting for a charger” into Real-time dispatch8 → 3 minLive in 28 cities100k+ users99% uptimeAI-driven decisionsAI-driven decisions.
The client operates a battery swap business with existing cabinets, battery inventory, and a market operations team. As the network expanded to 28 cities, devices became isolated — state could not sync in real time, the user flow grew complex, and peak loads destabilized the system.
Scan → open → swap → close looks like four steps. Underneath, the system handles battery identification, SoC / SoH sync, cabinet door control, order creation, billing, and cloud telemetry — all at once.
EV battery swap is a real-time infrastructure problem. It spans users, batteries, IoT devices, payments, ops, analytics, and city-level operations all at the same time.
Any single failure cascades straight into the user experience.
At peak: command latency, battery state drift, and unresponsive cabinet doors.
The business loop never closes end-to-end.
Translate the business problem into a system problem, then into a product shape. That sequence is how Wavesteam approaches every project.
Map the real chain, find the bottlenecks. Decide where things are stuck before deciding what to build.
Abstract devices, users, orders, and money flow into a unified event model — the shared language for every downstream system.
MVP → launch → retro → iterate, with blue-green deploys, canary releases, and observability baked in.
We walked a real (anonymized) user through the full swap flow and mapped each step to an actual mini program screen.
Backed by a private comms protocol, Kafka queues, async command scheduling, and redundant state checks, the platform sustains hundreds of thousands of daily swap requests.
“We used to manage devices. Now the system runs like a city-scale energy network.”
Each chain maps to a real subsystem. The cards below pair the UI, live metrics, and the design decisions that mattered.
IoT gateway + private protocol + MQTT / TCP keeps swap cabinets online, syncs battery state, dispatches commands in real time, and tracks device health.
Real-time sync of SoC, SoH, lifecycle, and charge/discharge state, feeding health management, risk detection, lifecycle analysis, and anomaly alerts.
The platform doubles as a business mid-platform: built-in CRM, store, campaigns, distribution, and tiered permissions support user ops, growth loops, commission settlement, and city-level expansion.
Kafka + ClickHouse + Airflow + BI power city-level ops analytics, user behavior analysis, battery dispatch prediction, and peak-load swap forecasting.
Automated alerting, honeypots and security scans, incident response, blue-green deploys, and canary releases drove > 99% uptime, with average incident time dropping from 4 hours to 16 minutes.
Geo trace for every vehicle, battery, and cabinet, with sub-second location for offline devices and auto-dispatched work orders.
IoT and data sit at the base. Business and algorithms close the loop in the middle. The application layer simply hands capability to operators and users.
Once the platform shipped, scale, system stability, and user experience all moved up together.
“We used to manage devices.
Now the system runs like a city-scale energy network.
Related cases
Other projects delivered by Wavesteam — AI, IoT, platform builds and enterprise software.
An operations platform for residential and online communities — group management, member workflows and dashboard visualization in one place.
View case
An in-vehicle BMS monitoring platform with a full data pipeline and a clean migration from a legacy system to a modern stack.
View case
A CAD-driven AI quotation system that collapses traditional mold pricing from days to minutes for manufacturing teams.
View caseThese projects usually involve business systems, device integration, AI workflows or multi-role back-offices. We assess feasibility against real delivery constraints and give recommendations close to the implementation stage.
