We are looking for a Python Engineer with strong RAG/LLM experience and supporting Node.js/React skills to help build and evolve a fast-moving internal platform and demo toolchain.

This is a hands-on, fullstack-oriented role focused on:

• Backend engineering in Python (FastAPI)
• End-to-end RAG pipeline development
• Document processing and AI integration
• Supporting frontend/backend integration (Node.js + React)
• Rapid prototyping and iteration in a POC-driven environment

Key Responsibilities

1. Backend & Platform Engineering

• Design and develop scalable APIs using FastAPI
• Implement asynchronous processing (Celery, messaging patterns)
• Design data models using PostgreSQL / MongoDB
• Build event-driven architectures (RabbitMQ, Azure Service Bus)
• Ensure clean architecture, strong typing, and maintainability

2. RAG / LLM Engineering

• Build end-to-end RAG pipelines:
  • ingestion → chunking → embeddings → retrieval → generation → structured output
• Implement hybrid retrieval:
  • vector search (embeddings)
  • BM25 sparse retrieval
  • metadata-based filtering
• Integrate with:
  • Azure AI Search (ACS)
  • Qdrant / pgVector
• Develop and maintain:
  • Prompt engineering workflows
  • Versioned prompts and guardrails
  • Structured output parsing (OpenAI APIs)

3. Document Processing & Knowledge Ingestion

• Build pipelines for legal/financial documents:
  • PDF extraction (PyMuPDF / fitz)
  • OCR for scanned files (Tesseract / pytesseract)
• Implement advanced chunking strategies:
  • clause-aware / section-aware segmentation
• Optimize ingestion for retrieval quality and accuracy

4. Fullstack Integration

• Maintain and extend Node.js + React services
• Ensure smooth integration between frontend and Python APIs
• Support backend components in Node.js when needed
• Maintain end-to-end system reliability

5. Internal Tools & Demo Development

• Build Streamlit-based internal tools:
  • multi-page apps
  • session state handling
  • interactive grids (streamlit-aggrid)
• Enable fast prototyping for:
  • demos
  • SME workflows
  • stakeholder validation

6. Data & Reporting

• Generate structured outputs:
  • Excel / Word reports (openpyxl, pandas)
• Support batch processing across multiple entities/projects

7. Quality & Delivery

• Write and maintain unit & integration tests
• Containerize services using Docker / Docker Compose
• Maintain reproducible configurations (YAML-driven)
• Follow strong Git practices (clean commits, PRs, CI/CD)

Required Skills (Must-Have)

Python & Core Engineering

• Python 3.12+, strong OOP and clean architecture
• CLI scripting (argparse, bash)
• Package management, virtual environments, Docker

Backend Development

• FastAPI (API design, service-oriented architecture)
• Asynchronous systems (Celery, messaging, event-driven design)
• PostgreSQL (or other relational databases)

RAG / LLM Engineering

• Hands-on experience with RAG pipelines
• Prompt engineering with versioning
• OpenAI API (including structured outputs)
• Embeddings + retrieval tuning
• Hybrid retrieval (vector + BM25)

Vector Search & Retrieval

• Qdrant (or similar: Pinecone, Weaviate, Milvus)
• Azure AI Search / Azure Cognitive Search

Document Processing

• PyMuPDF (fitz)
• Tesseract / pytesseract
• Advanced chunking strategies for structured documents

Internal Tools

• Streamlit (multi-page apps, session state)
• Interactive UI components (streamlit-aggrid)

Fullstack Support

• Basic–intermediate Node.js + React

Soft Skills

• Strong communication with business stakeholders (e.g., accounting/audit SMEs)
• Ability to work in rapid iteration / POC environments
• End-to-end debugging mindset (pipeline ownership)
• Analytical thinking and problem-solving

Nice to Have

• RAG evaluation frameworks and quality metrics
• LangChain, LangGraph, CrewAI, AutoGen, MCP/FastMCP
• Azure CI/CD and cloud-native deployments

Project Details

• Start Date: ASAP
• Duration: 12 months
• Location: Fully remote
  • Occasional in-person sessions (~1x per quarter, Prague)
  • Candidates based in Prague are a plus
• Working Hours:
  • Required overlap: 2–6 PM CET
  • Preferred: 2–7 PM CET
• Hiring Process: HackerRank challenge included

‍