Case study: Logistics platform

Re-architecting BookMyCargo's B2B logistics platform for multi-country scale

BookMyCargo, a B2B logistics platform, had to support high-volume, multi-country operations, distributed vendors and real-time coordination.

The situation

A B2B logistics platform had grown from a single-market product into one that had to coordinate shipments across several countries, many vendors and real-time status updates. The original design assumed one market, a handful of partners and a team in one place.

As volume grew, the platform had to handle more shipments, more integrations and more exceptions at the same time, while a distributed team kept adding features.

What was breaking

The pressure showed up in the same places it does in most growing logistics products:

  • Tightly connected components, so a slow partner integration could hold up unrelated work
  • Exceptions (delays, mismatches, missing documents) handled by people, one by one
  • Business rules buried in code, so every rule change needed a release
  • Releases that depended on a few people and were hard to coordinate across time zones

What I did

As Head of Architecture at BookMyCargo since 2018, I re-architected the platform as a cloud-native, event-driven system on AWS. Instead of each part calling the next and waiting, parts of the system publish events (a shipment created, a status changed, a document received) and the services that care about them react independently. A slow partner no longer holds up everything else.

Business rules moved into rule engines, so operations teams could change routing and validation rules without waiting for a release. AI-assisted exception handling sorted and suggested fixes for common problems, leaving people to decide on the unusual cases.

Behind the automation sit data pipelines and structured data models that feed operational reporting, so the business can see where shipments slow down and which rules to change next.

For the distributed team, I set CI/CD governance: automated build, test and deployment pipelines, release orchestration and clear ownership, so any team could release safely in its own time zone. Alongside the architecture work came performance and cost-efficiency programmes, planned against where the business was heading rather than where it was.

The result

High-throughput shipment processing with far fewer manual steps, and predictable releases across borders.

  • High-throughput shipment processing with far fewer manual steps
  • Predictable releases across borders and time zones
  • Rule changes made by operations without a code release
  • Partner integrations isolated, so one slow partner does not slow the platform
  • Operational reporting built on the same event data, instead of manual spreadsheets

What other teams can take from this

  • Event-driven design pays off when many independent parties change state at different speeds.
  • Moving rules out of code is often the fastest way to cut manual work.
  • AI is most useful on exceptions, where it sorts and suggests, while people keep the final say.

Services behind this work

MVP to Production Rescue

Project based

For products that proved the idea but cannot carry the growth. I turn the MVP into a system that is safe to build on.

  • Re-architecture of the weak parts, not a rewrite
  • CI/CD, release standards and observability
  • Database and performance tuning
  • Cloud cost reduction on AWS and GCP
MVP to Production Rescue: details

AI & Agent Readiness

Assessment or hands-on build

An AI readiness assessment, or a hands-on build, to bring AI agents into your product and operations without handing them the keys to production.

  • Where AI and automation will actually pay off
  • Least-privilege access and approval gates for agents
  • Audit trails that satisfy SOC 2 and HIPAA reviews
  • Agentic workflows designed end to end
AI & Agent Readiness: details

Tell me what is breaking.

I will tell you honestly whether I can fix it.