Avaronn, connected vehicles · 2025

Three vehicle apps, one event stream, and coordination done by hand

Led by Ansuj Joshi as co-founder and AI tech lead at Avaronn, Jan to Dec 2025 (exited 2025)

Situation

A connected-vehicle ecosystem with three production apps needed real-time telemetry and assignments that were being coordinated manually.

What we built

A Kafka event-driven backend with real-time telemetry and automated assignment workflows.

What changed

Manual coordination fell by 70%.

The problem

Vehicles, drivers and jobs produced events that three different apps needed in different shapes, and people in the middle were relaying them.

What we built

  • A Kafka event backbone that every app publishes to and subscribes from.
  • Real-time telemetry ingestion and automated assignment workflows that replaced the manual relay.
  • Code reviews, testing, debugging and CI/CD pipelines, with ML integrated across the three apps.

How it works

  • Events are the source of truth; each app projects the stream into the view it needs.
  • Assignment rules run on the stream, so a new job is matched the moment it appears.

What changed

70%

less manual coordination

Source: Ansuj Joshi resume, 2026

Stack

KafkaPythonKubernetesAWS

If this looks like your situation

Tell us where it hurts. We will say whether it is worth automating, and what we would build first.