Services / Overview
Focused on what eats time in practice: recurring Excel work, reports, data prep, and grown Office systems. Each area has its own service page.
Office automation as one discipline: Excel macros with Option Explicit and custom error classes, Access apps with 32→64-bit migration, backend moves to SQL Server. Refactor beats rebuild.
Learn more
Template architecture instead of spreadsheet chaos: separated input/logic/output zones, ListObjects instead of range addresses, validation, protection logic, a file your team still understands 18 months later.
Learn more
Monthly and weekly reports from click marathon to a reproducible pipeline: Power Query as ETL, Power Pivot/DAX as model, KPI definition table, audit trail on every refresh.
Learn more
Lean internal helpers for recurring tasks: input masks, import pipelines with validation, PDF and email outputs on fixed schemas. Small scope, clear ownership, no enterprise overhead.
Learn more
Structured team knowledge: glossaries, process documentation, handover skeletons, searchable snippets. Head knowledge becomes team knowledge, without anyone needing wiki discipline first.
Learn more
Office and file workflows as one strung-through pipeline: CSV imports with schema validation, PDF chains with reproducible layout, Outlook chains with logging. Where Excel hits limits, Power Query and Python step in.
Learn more
Reproducible ETL with the M language: typed imports, query folding back to source, parameter tables, a custom-function library, schema validation with Try/Otherwise. Refresh often 5-30x faster.
Learn more
Where Excel hits limits: typed Python pipelines with pandas/polars, SQLAlchemy, pydantic, pytest, loguru. SQL tuning via execution plans, views, materialized views, partitioning.
Learn more
Local LLMs (Ollama, vLLM, llama.cpp), RAG with Qdrant/pgvector, n8n as on-prem orchestrator, MCP servers as a tool bridge. GDPR and EU AI Act as part of the architecture, not bolted on.
Learn more
The VBA discipline covers Excel and Access together because a senior engineer treats both as one toolset. AI has its own page because the topic needs more room than a service card.
/ Next step
A 30-minute intro call costs nothing. You outline the topic, I give an honest read, directly with the engineer, no sales sequence.