Software Architect

About GoComet

GoComet is a global AI-powered logistics and supply chain platform transforming how businesses manage end-to-end freight operations. Trusted by 500+ enterprises across 70+ countries, our award-winning solutions include:

• GoProcure – Automated RFQ management with dynamic vendor bidding, audit-ready bidding logs, and cost-optimized freight procurement.
• GoTrack – Real-time container tracking with predictive ETAs, geofencing, and proactive alerts across multiple carriers.
• GoInvoice – Intelligent invoice reconciliation using OCR and NLP to reconcile invoices with quotes and efficiently handle discrepancies.
• Logistics Control Tower, Smart Schedules, Port Congestion Tools, Freight Indexing, and Data Analytics – delivering complete visibility, market insights, optimized planning, and workflow automation.

Featured in Forbes, Gartner, and Logistics Tech Outlook, GoComet enables greater transparency, automation, and cost savings across multimodal logistics.

Role: Architect / Software Engineer

The Role

You own the platform, end to end.

You will design the connector framework and then write it — the auth and tenancy model, event backbone, sync engine, tool surfaces an agent calls, deployment path, and the trace you read when a client's data stops arriving at 02:00.

There is no platform team to hand it to.

You will make the calls that are expensive to reverse — scalability, security, isolation, and what belongs to a connector versus a gateway — and make them early, with incomplete information, alongside people who will push back.

1. The Connector Framework

Shape, codegen, typed clients, versioning — how the fiftieth integration costs a fraction of the first.

2. Auth & Tenancy

Per-client credentials, rotation, scopes, and the isolation boundary that cannot leak.

3. Event Infrastructure

Webhooks, queues, retries, idempotency, replay, and ordering guarantees.

4. Sync Engine

Incremental and full syncs, backfills against live traffic, and reconciliation.

5. The Agent Edge

Tool schemas, MCP surfaces, what an agent may call, and what it may never do.

6. Reliability & Scale

Budgets, rate limits, observability, and the trace you read at 02:00.

What We're Looking For

We weight systems you have built over tools you have used.

Framework fluency is widely available. What is scarce is an engineer who has designed the boundaries, shipped them, and carried responsibility for how they perform under load.

• You have built a platform, not only used one.
• You chose the boundaries and data model, ran it in production, and know which decisions you would make differently today.
• Distributed systems are your home ground.
• APIs, queues, consistency, partial failure. You have debugged the thing that only breaks under concurrency — and you write the postmortem yourself.
• You are hands-on and intend to remain so.
• This is an architect who writes code. There is no separate team to translate a design into a system; both are yours.
• You have integrated a genuinely difficult system.
• An undocumented enterprise API, a partner that changes payloads without notice, or a system with no sandbox. You made it dependable regardless — and can explain how.
• You test your own designs for weaknesses.
• Building a platform from first principles means actively disconfirming your own design. Expect the interview to examine yours closely.

Nice to Have — Genuinely Optional

• Prior integration-platform experience
• AI or LLM exposure — useful, but readily learned here
• Logistics knowledge — not expected; we would rather teach the domain to a strong systems engineer
• A specific degree — not a requirement