Principal Engineer (Founding) Insurance Quoting Automation

Prestige Trucking Insurance is a US commercial trucking insurance agency, hiring our founding senior engineer in India. We've built Insure Stack: upload a single intake form, and it quotes every major trucking carrier in parallel turning a process that takes an agent hours into one that should take seconds. The product is live, has real users, and has real, hard problems. You would be one of the first engineers on the team, working directly with our part-time CTO (ex-Meta).

The role :

You are the senior-most full-time engineer in the company. You own architecture, technical direction, and the standard of the codebase and you are still writing Python every week. This is not a management role and will not be one for at least a year.

The honest version of the problem :

Our target is all carriers quoted in parallel in under 10 seconds. Today, when a quote fails, a human takes 12 hours to recover it by hand, often introducing errors. That is a very large gap, and we believe it is architectural, not a matter of effort. We want someone who reads that and immediately starts forming hypotheses: sequential calls where there should be fan-out, cold browser contexts and re-login on every request, no per-carrier timeout budget, no partial-result path, no trace artifacts so every failure is re-diagnosed from scratch, manual fixes applied in the UI and never committed back. You come in, work out which of those it actually is, and fix it.

What you'll own :

- End-to-end architecture of the quoting pipeline (concurrency model, timeout budgets, partial-result semantics, failure isolation)

- The reliability bar and the observability to prove it

- Technical direction and code review for a small team alongside our CTO

- Build-vs-buy and infrastructure decisions

- The hiring bar for future engineers

What we need :

- 8+ years building production backend systems, with real time at senior / staff / principal level

- Deep Python you have made async Python behave correctly under real concurrency, not just used the async keyword

- You have owned a system where latency was the product requirement, and can talk specifically about how you found and removed the time

- Experience designing for partial failure: timeouts, retries with backoff, circuit breakers, idempotency, graceful degradation

- You have worked somewhere small enough that you owned production outcomes, not just your tickets

- Clear written communication; the team is distributed

Strong pluses :

- Regulated-industry background (insurance, fintech, healthcare) where wrong output has consequences

- Any experience with browser automation, scraping, or systems that depend on third-party interfaces you don't control

- Having inherited a struggling system and made it reliable

- Insurance / logistics / trucking domain knowledge

What this is not :

This is not distributed-systems-at-hyperscale. Our problem isn't millions of requests per second it's a handful of flaky third-party websites that change without warning and must be handled correctly every time. If your strongest work is horizontal scaling, you may be excellent and still not the fit here.