Location: Warsaw, Poland - Hybrid
Capco Poland is a leading global technology and management consultancy, dedicated to driving digital transformation across the financial services industry. Our passion lies in helping our clients navigate the complexities of the financial world, and our expertise spans banking and payments, capital markets, wealth, and asset management. We pride ourselves on maintaining a nimble, agile, and entrepreneurial culture, and we are committed to growing our business by hiring top talent.
ROLE OVERWIEW:
This role is responsible for designing and implementing ingestion pipelines, document processing workflows, data normalization, and integrations across Azure services (SharePoint, Microsoft Graph, Azure AI Search, Azure Document Intelligence, App Insights). The engineer will also implement automated evaluation frameworks and quality monitoring mechanisms.
The primary objective of this role is to enable domain agents, task agents, and the master agent to operate on properly prepared, complete, indexed, and secure knowledge artifacts, in full alignment with enterprise security, compliance, and banking standards.
Key ResponsibilitiesDesign and develop data connectors (SharePoint, OneDrive, Microsoft Graph, GCP sources, external APIs) to enable automated document and metadata ingestion.
Implement ingestion pipelines (batch and event-driven) and processing orchestration mechanisms.
Process documents using Azure Document Intelligence, including parsing, OCR, layout extraction, structured data extraction (tables, fields, confidence scoring).
Develop transformation and normalization workflows, including data cleaning, segmentation, PII masking, and generation of structured knowledge artifacts.
Index content into Azure AI Search and Knowledge Bases, including index design, indexers, skillsets, enrichment pipelines, embeddings, and vector stores.
Prepare evaluation datasets (baseline, ground truth, domain-specific test cases).
Automate quality evaluation of extraction and indexing processes (precision, recall, fidelity metrics, drift detection).
Implement instrumentation, logging, and monitoring using Application Insights and Log Analytics.
Optimize document processing costs (batch vs. on-demand processing strategies, layered caching, cost-per-document analysis).
Establish CI/CD pipelines and development standards (GitHub Actions, testing, code quality, linting, artifact registry, containerization).
Build and maintain containerized services (Docker, Azure Container Registry, Azure Container Apps / AKS / WebApp for Containers).
Automate ingestion and evaluation workflows using Azure Functions, Logic Apps, and Durable Functions.
Strong Python skills (asyncio, FastAPI, Pydantic, multiprocessing).
Hands-on experience with Azure SDK for Python (Storage, Cognitive Services, AI Search, Application Insights).
Practical experience with Microsoft Graph API / SharePoint API (file and metadata retrieval).
Azure AI Search: index design, indexers, skillsets, embeddings, vector search.
Azure Document Intelligence (OCR, layout extraction, custom models).
Testing experience (pytest, integration tests, cloud service mocking).
GitHub Actions (build, test, scanning, artifact management, deployment).
Docker (image building, multi-stage builds, layer optimization).
Logging and monitoring (Application Insights, Log Analytics).
Experience working with large document collections (batch ingestion at scale).
Data security best practices (PII masking, RBAC, Entra ID integration).
MCP Client / MCP Tooling (custom agent integrations).
Experience with LangChain or Semantic Kernel (RAG / agent pipelines).
GCP experience (BigQuery, Looker, Cloud Functions) within a multi-cloud Knowledge Management architecture.
AKS / Azure Container Apps / WebApp for Containers for scalable ingestion services.
Durable Functions for orchestrating large document processing workflows.
We offer a flexible collaboration model based on a B2B contract, with the opportunity to work on diverse projects.
