Senior Backend Engineer (FastAPI)
⚲ Warszawa
30 420–47 531 zł netto (+ VAT) / mies.
Wymagania
- Python
- FastAPI
- PostgreSQL
- SQL
- Redis
- Celery
- ARQ
- LangGraph
- RabbitMQ
Opis stanowiska
Nasze wymagania:
Seven years or more delivering production backends — though how deep you go matters more to us than the figure itself
You think in failure modes, not just the happy path
You've shipped genuine LLM systems (RAG, tool-calling, agents), or you can't wait to — and you've got a feel for where they fall apart
You lean on AI coding tools (Claude Code, Cursor, Copilot) and get real value from them: more speed, yes, but you grasp and back every line — code you'd reject in someone else's review is code you won't merge in your own
A startup that keeps changing course doesn't faze you — plans get set, then torn up and redrawn
Mile widziane:
RAG or vector search experience
LangChain or LlamaIndex
O projekcie:
You'll lead the backend that drives the AI features — the APIs, the data pipelines, and the orchestration that connects the product to the language models doing the actual work. In practice that's querying across large document sets, getting trustworthy answers out of LLMs and their tool calls, pushing results to the UI as they stream, and grinding through lengthy background jobs without the whole thing toppling over.
Forget CRUD. The real difficulty is making AI systems fast and reliable on confidential data, at production scale. The architecture choices land on you early, and you'll be the one living with them.
Tech stack experience
• Python: 6+ years
• FastAPI: 3+ years
• PostgreSQL / SQL
• Redis: 2+ years
Zakres obowiązków:
Architect and ship the APIs and services that everything else depends on
Own the AI plumbing end to end: retrieval, orchestration, tool-calling, streaming, and the evals that tell you whether the results are worth anything
Make async and distributed workloads rock-solid — document ingestion, embeddings, and long jobs running on Redis
Wire up integrations with what clients already use (email, document stores) and harden them for real-world traffic
Put data security and isolation front and center, because clients are handing over material they can't afford to leak
Seven years or more delivering production backends — though how deep you go matters more to us than the figure itself
You think in failure modes, not just the happy path
You've shipped genuine LLM systems (RAG, tool-calling, agents), or you can't wait to — and you've got a feel for where they fall apart
You lean on AI coding tools (Claude Code, Cursor, Copilot) and get real value from them: more speed, yes, but you grasp and back every line — code you'd reject in someone else's review is code you won't merge in your own
A startup that keeps changing course doesn't faze you — plans get set, then torn up and redrawn
Mile widziane:
RAG or vector search experience
LangChain or LlamaIndex
O projekcie:
You'll lead the backend that drives the AI features — the APIs, the data pipelines, and the orchestration that connects the product to the language models doing the actual work. In practice that's querying across large document sets, getting trustworthy answers out of LLMs and their tool calls, pushing results to the UI as they stream, and grinding through lengthy background jobs without the whole thing toppling over.
Forget CRUD. The real difficulty is making AI systems fast and reliable on confidential data, at production scale. The architecture choices land on you early, and you'll be the one living with them.
Tech stack experience
• Python: 6+ years
• FastAPI: 3+ years
• PostgreSQL / SQL
• Redis: 2+ years
Zakres obowiązków:
Architect and ship the APIs and services that everything else depends on
Own the AI plumbing end to end: retrieval, orchestration, tool-calling, streaming, and the evals that tell you whether the results are worth anything
Make async and distributed workloads rock-solid — document ingestion, embeddings, and long jobs running on Redis
Wire up integrations with what clients already use (email, document stores) and harden them for real-world traffic
Put data security and isolation front and center, because clients are handing over material they can't afford to leak
🔍 Dekoder Ogłoszenia
🟡
Seven years or more delivering production backends — though how deep you go matters more to us than the figure itself
Chociaż wymagają 7 lat doświadczenia, faktycznie szukają osoby z głęboką wiedzą, która może być osiągnięta w krótszym czasie.
🟢
You think in failure modes, not just the happy path
Oczekuje się, że będziesz proaktywnie identyfikować i rozwiązywać potencjalne problemy, a nie tylko skupiać się na działaniu w idealnych warunkach.
🟡
You've shipped genuine LLM systems (RAG, tool-calling, agents), or you can't wait to — and you've got a feel for where they fall apart
Poszukują kogoś z praktycznym doświadczeniem w systemach LLM, lub kogoś, kto jest bardzo chętny do nauki i rozumie ich ograniczenia.
🔴
A startup that keeps changing course doesn't faze you — plans get set, then torn up and redrawn
Oczekuje się od kandydata dużej elastyczności i odporności na częste zmiany priorytetów i kierunku rozwoju projektu.
🟢
Forget CRUD. The real difficulty is making AI systems fast and reliable on confidential data, at production scale.
Oczekują, że będziesz rozwiązywać złożone problemy związane z wydajnością i niezawodnością systemów AI na dużą skalę, a nie tylko standardowe operacje na danych.