Most founders evaluate engineering studios the wrong way. They look at the portfolio, ask for a quote, and pick whoever's cheapest. Then six months later they're rewriting half the codebase. This is the checklist we wish more founders used — written from the inside, after building 33+ platforms at Xenotix Labs.

1. Ask to see one full architecture document

Not a pitch deck. Not a portfolio. The actual architecture document for a real project. Microservices boundaries, data model, event flows, deployment topology, scaling assumptions. If they can't show you one (NDA-redacted is fine), they don't write them. If they don't write them, they don't think systematically about scale.

At Xenotix Labs we share redacted architecture docs from D2C dairy commerce (Veda Milk), real-time cricket trading (Cricket Winner), legal-tech super-apps (Legal Owl), and offline-first rural education (7S Samiti). You can read public versions in the engineering blog.

2. Ask how they handle idempotency on the first call

This is a trap question. If their answer is "we use idempotency keys on critical endpoints" — good. If their answer is "what do you mean by idempotency?" — run.

Idempotency is the difference between an engineering team that knows distributed systems and one that's about to double-charge your customers in production.

3. Ask what they'd do differently if they rebuilt their last project

If the answer is "nothing, it was perfect" — they haven't run it in production long enough to learn anything. Pass.

If the answer is a 10-minute story about a tradeoff that aged badly — hire. Engineering scars are the only proof of engineering experience.

4. Ask to see the admin panel of a shipped project

User apps are easy to make pretty. Admin panels reveal whether the studio actually thinks about operations — payouts, refunds, dispute resolution, user verification, content moderation, analytics. A bad admin panel makes the operations team your bottleneck after launch.

We build admin panels as first-class Next.js apps with role-based access, audit logs, server-side filtering, and tight integration with our background-job system. The admin panel is where the business actually lives.

5. Ask about the testing pyramid

Unit tests, integration tests, end-to-end smoke tests — they should have all three. If they say "we test in production" or "our QA team does manual testing" — you're going to get 2 a.m. incidents in your first month live.

6. Ask about offline-first if your users are in India

70% of Indian users hit dead zones daily (basements, gated societies, rural areas, tunnels). If you're building anything for delivery boys, field surveyors, or rural users — you need offline-first Flutter with local SQLite, sync workers, and conflict resolution. Most studios skip this. The ones who don't, ship apps that work where users actually use them.

7. Ask about deployment and CI/CD

"We deploy to AWS" is not an answer. Ask: how long is a deploy? How do you do zero-downtime? What's your rollback strategy? Do you have staging that mirrors production? How do you handle database migrations on tables with 1M+ rows?

If those answers are vague, every deploy is going to be a stressful event. If they're crisp, deploys become a non-event.

8. Ask for the actual rate, in writing, with what's included

Vague pricing means cost overruns later. Ask: hourly or fixed bid? What's included in the rate (design, mobile, web, backend, DevOps)? How are change requests handled? What happens if a deadline slips? Get it in writing before you sign.

At Xenotix Labs our pricing is clear: per-project fixed bid for MVPs (4-12 weeks), hourly retainer for ongoing work, with monthly invoices and audit trails. No surprise change orders.

9. Ask who actually writes the code

Many studios subcontract or use juniors for actual implementation while senior architects only attend kickoff calls. Ask: who exactly will work on my project, week by week? Can I talk to them directly? What's their experience level?

If you can't get those answers, you don't know who's building your product.

10. Ask for a 30-min architecture review of YOUR idea

This is the single most useful filter. Bring your business idea and ask for a free 30-minute architecture review. A studio that can sketch a sane architecture in 30 minutes will sketch a great one in 30 hours. A studio that's vague or sales-pitch-heavy is going to be vague in implementation too.

We offer these for free at https://xenotixlabs.com — it's the best way to evaluate fit before signing anything.

Final advice

The cheapest engineering studio is rarely the cheapest engineering studio. Rework, rewrites, missed deadlines, and operational pain are the real costs. Hire for engineering discipline, ask hard questions, and look at admin panels, not landing pages.

If you're building a D2C product, marketplace, AI-powered tool, or any platform that needs to scale — Xenotix Labs is happy to be one of the studios you evaluate. Even if you don't pick us, the 30-min review will sharpen your thinking.