5 October, version 1 — Senior Python Backend Engineer, digital healthcare
Prep for the first tech check: a long-term senior backend role on an
international AI healthcare platform operating at large scale. The raw ad is
in description.txt, and the resume both interviewers saw is the .docx
beside it; both stay local and are never published.
Version 2 is the second interview — a similar ad with a different emphasis. It links back here for every answer that applies unchanged.
Same conventions as the other tech checks — the checklist of Company A and Company B, and the spoken answers of 14 August.
What kind of interview this is
A generalist senior backend ad. It names no framework, no cloud and no AI requirement, so expect a breadth check at senior depth — Python, a web framework, REST, Postgres, integrations, testing — with healthcare as the context your answers should be framed in.
The responsibilities list is the tell. Owning work from design through to production, architectural decisions, code reviews and mentoring describe someone who raises the team. Expect each of them back as a behavioural question, and expect “why” and “what did it cost” after every technical answer.
The ad, decoded
| Ad line | What they will actually probe | Weight |
|---|---|---|
| 5+ years of commercial Python | GIL and free-threading, async vs threads vs processes, memory, the data model | 🔴 |
| Modern Python web frameworks | FastAPI in depth — DI, async def vs def, testing; Django ORM if that is their core |
🔴 |
| REST APIs | resource design, status codes, idempotency, pagination, versioning, errors, auth | 🔴 |
| Backend architecture, scalable services | layering, monolith vs microservices, caching, queues, scaling a stateless service | 🔴 |
| Databases, migrations, query optimisation | EXPLAIN (ANALYZE, BUFFERS), indexes, N+1, locking, zero-downtime migrations |
🔴 |
| Integrate internal and third-party services | timeouts, retries, idempotency keys, circuit breakers, webhooks, the outbox | 🔴 |
| Unit and integration tests | what to mock and what not, real Postgres in tests, dependency_overrides |
🔴 |
| Architectural decisions and code reviews | trade-off reasoning, ADRs, what you look for in review and how you say it | 🟡 |
| Support production services | observability, incident response, postmortems, safe deploys | 🟡 |
| Mentor less experienced engineers | one mentoring story with a result you could see | 🟡 |
| Healthcare, international, large patient volumes | PHI, HIPAA and GDPR, audit logging, FHIR, data residency | 🟡 |
| Distributed team, English Upper-Intermediate+ | reasoning aloud in English; written, asynchronous habits | 🟡 |
| A platform using AI | a model call treated like any unreliable dependency, under PHI rules | 🟢 |
Your best evidence
Both interviewers saw the same resume, and you will talk about its two most recent projects. Between them they cover most of this ad:
- Automation Platform (Jan–Sep 2026) — the most recent, so the one they ask about first. Serverless Python for underwriting workflows: Lambda functions orchestrated by Step Functions, state in DynamoDB behind API Gateway, AI extraction, a multi-stage rules engine, REST integrations with Salesforce and OAuth2 services — Microsoft Graph and Slack — and two Streamlit apps on ECS with Cognito SSO. Your integrations and data processing story.
- Invest Planning Platform (Sep 2025 – Jan 2026) — FastAPI services and REST APIs, PostgreSQL and SQLAlchemy models including audit and consent records for a regulated product, third-party integrations, TDD with pytest, Sentry. Your regulated-domain story — the closest thing to healthcare experience without having worked in healthcare. Frame it that way.
What neither shows is the second half of the responsibilities: architecture decisions, production incidents and mentoring — and the resume mentions code review only on an older project. Those need one story each; the skeletons are in The call — everything except the technical answers.
The files
| File | What it is | When |
|---|---|---|
| 5 October — Senior Python Backend Engineer — prep map | the checklist — every line of the ad, linked to the notes that answer it, plus the design question to rehearse | the days before |
| Answers — say these out loud | 23 likely questions, answered the way you would say them | the day before, aloud |
| The call — everything except the technical answers | self-intro, the project story, the behavioural questions, your questions, the close | the day before, aloud |
| Last hour — read this, then stop | one page of one-liners, tables, dated facts and your own numbers | the hour before |
Refresh plan, by the time you have
A day or more. Work down 5 October — Senior Python Backend Engineer — prep map, the 🔴 sections
first. Open the bold note for anything you cannot explain aloud in a minute,
and tick only what you could defend under a follow-up. Then fill every
[bracket] in 02 and 03 with real facts and numbers, and say both files
aloud once.
Three hours. Skip the map. Read Answers — say these out loud aloud, timing each answer at 60 to 90 seconds. Then the self-intro and the project story in The call — everything except the technical answers, three times each. Then walk the design question at the end of 5 October — Senior Python Backend Engineer — prep map once.
One hour. Last hour — read this, then stop, then the self-intro twice, aloud. Nothing new.
Ten minutes. The one-liners and your numbers in Last hour — read this, then stop, and the three rules below.
The three things that decide it
- Reason aloud. Context, options, decision, trade-off, result — 60 to 90 seconds per answer. A correct fifteen-second answer reads as mid-level.
- Bring specifics. A number, a query plan, an incident, a decision and what it cost. Senior is heard as specifics, not as coverage.
- Speak the domain. One sentence each on PHI in logs, audit logging as a requirement, BAAs and FHIR R4 puts you ahead of every candidate who has only generic backend answers.