Condense in Action: How a Tyre OEM Built and Scaled Its Fleet Platform Using Real-Time Streaming with Condense

From zero real-time infrastructure to a fully operational, scalable Fleet Management System in 4 months, powered by Condense, deployed on their own cloud (Microsoft Azure), and operated without an in-house streaming team. 

Description 

A leading commercial tyre manufacturer ventured into unfamiliar terrain: building a Transport Management System (TMS) for digital fleet operations. With no prior experience in cloud-native pipelines, event-driven architectures, or telematics integration, their new software team was starting from scratch. 

What followed was a technical co-creation journey. With Zeliot’s domain guidance and the Condense platform, the team deployed a production-grade TMS capable of ingesting, parsing, transforming, and operationalizing real-time data from thousands of heterogeneous vehicles across makes, protocols, and geographies. 

TLDR;

• Timeline: September 2023 (Initiation) → January 2024 (MVP launch) 

• Cloud: Customer's own infrastructure (BYOC deployment) 

• Platform: Kafka-native, Condense-managed real-time stack 

• Vehicles onboarded: 4000+ (as of June 2025) 

• Growth rate: 200+ vehicles added monthly 

• Business velocity: 6× faster GTM cycle per feature 

Highlights 

• Clean-slate architecture, built by a team with zero streaming background 

• End-to-end ingestion, transformation, enrichment, and delivery powered by Condense 

• Application team focused entirely on features; Condense handled all infra and FinOps 

• Achieved high-frequency feature releases without any scaling regressions 

• Multi-environment (Dev/UAT/Prod) deployments with no manual infra management 

Executive Summary 

This is not a typical "digital transformation" story. 

This is the story of a tyre manufacturing enterprise, deeply rooted in industrial operations, launching a digital business unit with a singular goal: build and operate a scalable Transport Management System. 

There was no inherited software platform, no Kafka stack, and no backend engineers with cloud-native experience. 

Instead, what they had was: 

• Clear business intent 

• A new product team 

• Access to vehicle data from fleets already using their tyres 

• And a partnership with Zeliot, who brought in Condense 

Condense abstracted the underlying complexity of real-time systems. In just four months, the team went from zero to a fully functional TMS, handling diverse telematics payloads, trip states, geofencing logic, and operational alerts. 

But the real value emerged after launch. With Condense managing uptime, patching, Kafka operations, throughput-based autoscaling, and multi-environment provisioning, the TMS team focused exclusively on: 

• Frontend logic 

• New feature pipelines 

• Customer onboarding workflows 

• Business-side telemetry visualization 

By mid-2025, the platform supports over 4000 vehicles and continues to add hundreds more per month, with no degradation in latency, throughput, or operational overhead. 

What the Engineer Manager has to say

About the Tyre OEM 

The client is a well-established commercial tyre manufacturer with deep relationships in India’s logistics and long-haul trucking ecosystem. In 2023, the company launched a new digital services business to provide Transport Management solutions to its customer base many of whom lacked access to reliable, scalable fleet intelligence systems. 

This business unit was built independently from their manufacturing operations, with a mandate to deliver a production-ready platform from scratch. 

How Did Condense Help? 

The Problem: How Do You Build a Real-Time System? 

The newly formed TMS team had deep knowledge of fleet behavior and logistics but no backend, no telemetry platform, and no Kafka experience. The goal was to ingest and act on real-time data from third-party and OEM telematics units across thousands of vehicles. 

The TMS team had no prior experience with: 

• Kafka 

• Stream processors 

• Cloud-native infra provisioning 

• Data schema evolution or versioning 

• Scaling ingestion pipelines with traffic bursts 

• Multi-protocol telematics decoding 

Even basic questions like how to parse a CAN bus event from one device and correlate it with a trip state had no internal answers. They needed a platform that already understood mobility and was already built for streaming. 

Condense filled that gap not with a toolkit, but with a full system. 

The Architecture That Made It Possible 

Condense was deployed in the manufacturer’s own cloud environment via its BYOC model, ensuring compliance, flexibility, and data control. Within days, the team had a working pipeline from raw vehicle data to live dashboard events. 

Below is a breakdown of how the solution was built technically, structurally, and operationally. 

1. Ingestion Layer: Multi-OEM, Multi-Device Handling 

• Connectors prebuilt for popular telematics vendors (AIS-140, OEM APIs, MQTT brokers) 

• Payloads ranged from nested JSON to binary and proprietary formats 

• Ingestion enriched with metadata (VIN, timestamps, location) at the source 

2. Parsing and Transformation Layer 

• Git-backed transforms deployed via Condense inbuilt IDE and CI pipelines 

• Event logic included trip segmentation, vehicle status classification, and alert tagging 

• JSON schema validation with fallback handling for malformed or unexpected fields 

• Streaming enrichment by joining vehicle master data and driver ID tables 

3. Delivery to Application Layer 

• APIs generated directly from topics and keyed event streams 

• Real-time WebSocket feeds used for live dashboards 

• Alert delivery via HTTP webhooks and Kafka topics to external systems 

• Kafka sink connectors for long-term archival into warehouse-compatible stores 

4. Multi-Environment Infra with Zero Manual Setup 

• Condense provisioned fully isolated clusters for Dev, UAT, and Production 

• Secrets management, autoscaling, and topic partitioning handled via configuration only 

• Version rollbacks, transform deployment history, and Git commit audit built-in 

5. FinOps and Autoscaling 

• Condense included usage-aware compute management: 
  → Automatically scaled vCPU and memory per topic volume 
  → Cluster size adjusted by vehicle count and ingestion patterns 

• Real-time cost dashboards surfaced usage by connector, transform, and delivery stream 

• Enabled precise cost forecasting per customer and per data source 

Summing all up  

Condense is a Kafka-native, industry-verticalized streaming platform designed to abstract operational complexity for teams building real-time applications. In this use case, what the team gained from Condense was not just a Kafka cluster. It was a complete data backbone: 

Without needing to manage clusters, debug brokers, or understand Kafka internals, the TMS team focused purely on: 

• Feature rollout (trip planning, geofence violations, route analysis) 

• Visual telemetry dashboards 

• Customer usage analytics 

• Quarterly roadmap execution 

By decoupling streaming infrastructure from application development, Condense allowed the TMS team to focus entirely on user features and market expansion, without worrying about the backend. 

Switch to Condense

If you’re a logistics platform, OEM, or telematics-backed service provider starting from scratch, you don’t need to build Kafka expertise to go live. 

Condense gives you the backbone so you can build the application. 

Book a meeting with us here to discuss your use-case and switch to a Real Time Data Streaming Platform, that just works!