Backend Developer

Position

Backend Developer

Reporting To

Solution Architect

Experience

3–4 years of professional backend development experience, primarily in Python

Must-Have Skills

Backend

- Python 3.11+ / 3.12+.
- FastAPI, with Uvicorn/ASGI — building and running async services in production.
- Strong REST API design sense — resource modelling, versioning, error handling, and pagination done properly, not just endpoints that work.
- Pydantic v2 for request/response validation and settings management.
- JWT authentication (RS256/JWKS) — we use Clerk for identity, so experience verifying externally-issued JWTs is important.
- A strong understanding of asynchronous backend development — you should know why you'd reach for async/await and where it can go wrong (blocking calls, event-loop starvation, etc.), not just how to write the syntax.

Database

- PostgreSQL, with a strong understanding of: transactions, locks, indexes, query performance, and tenant isolation.
- SQLAlchemy 2.0 ORM and Alembic migrations.
- Async PostgreSQL using asyncpg and raw SQL — no ORM. You should be comfortable writing and reasoning about SQL directly when the ORM isn't the right tool.
- PostgreSQL Row-Level Security (RLS) for enforcing tenant isolation at the database layer.

AI / LLM

- LLM application development using Anthropic Claude or OpenAI.
- Prompt design for reliable, repeatable outputs.
- Structured extraction, and tool/function calling with JSON output.
- RAG (retrieval-augmented generation) implementation, end to end.
- Vector databases — Qdrant, pgvector, or Pinecone.
- Embedding models — Voyage or OpenAI.

Document Processing

- PDF processing — pdfplumber or PyMuPDF.
- OCR — Tesseract or poppler.
- DOCX / XLSX / PPTX parsing.

Infrastructure

- Redis for caching, rate limiting,



and background job state.
- A solid understanding of background jobs and asynchronous processing more broadly.

Good to Have

- Multi-tenant authentication/tenancy design.
- Quotas and audit trails.
- Advanced RLS policy design beyond the basics.
- Background job cancellation and retry/recovery patterns.
- Prompt-injection defenses for LLM-facing endpoints.
- Experience building deterministic scoring engines where a model narrates results rather than computing them.
- Deployment on Railway; Docker / Nixpacks.
- Resend or a similar transactional email service.
- Financial, compliance, ESG, sustainability, or other regulated-reporting domain experience.

What You'll Do

- Design and build async FastAPI services, with real ownership over your area of the system.
- Work comfortably in both an ORM-based codebase (SQLAlchemy 2.0 + Alembic) and a raw-SQL, asyncpg-based codebase — and know when each approach is the right call.
- Design PostgreSQL schemas and RLS policies that keep tenant data properly isolated, and reason carefully about transactions, locking and query performance under load.
- Build and maintain LLM/RAG pipelines — prompt design, structured extraction, tool calling, and retrieval over vector stores — that respect our "deterministic core, AI narrates" architecture.
- Build document-processing pipelines that turn PDFs, scanned documents, and Office files into structured,



usable data.
- Use Redis for caching, rate-limiting and background job state, and build background jobs that handle retries and failure gracefully.
- Verify and work with Clerk-issued JWTs (RS256/JWKS) to secure APIs correctly.
- Review pull requests, raise the bar on code quality, and mentor trainee/junior engineers on the team.

Who Should Apply

- Bachelor's or Master's degree in Computer Science, Information Technology, Electronics, or a closely related engineering discipline.
- 3–4 years of professional backend development experience, primarily in Python, including real production use of FastAPI (or a comparable async framework).
- Demonstrated experience with PostgreSQL at a level beyond CRUD — you've had to reason about locks, indexes, or query plans under real load at some point.
- Some hands-on experience building an LLM-powered feature in production (not just experimenting with an API in a notebook) — prompting, structured output, or RAG.

Personal Competencies

- Strong ownership mindset — comfortable taking a backend feature from an ambiguous requirement to a shipped, production-ready result with minimal hand-holding.
- Able to give and receive direct technical feedback, and to mentor less experienced engineers without being asked twice.
- Comfortable moving between two different backend codebases (ORM-based and raw-SQL) without losing rigor in either.
- Clear written and verbal communication — able to explain technical trade-offs to both engineers and non-technical stakeholders.
- Good time management and the discipline to work independently as well as in a team.
- Adaptability — this is a fast-moving product team and priorities shift as the applications evolve.

📌 Desarrollador de back-end (Colombia)
🏢 Valenta
📍 Colombia