Data consulting without the fluff

I solve real data problems: from dashboards that don't add up to broken pipelines or ML models that never reach production. No upselling what you don't need.

Tell me your problem

Services

Power BI & DAX

Data models that load fast, dashboards that make sense, and DAX that doesn't break when someone adds a filter.

Who it's for

Analysts with slow reports, companies moving from Excel to real BI, teams with Power BI but nobody who actually knows DAX.

Includes

→ Data model optimization (VertiPaq)
→ Documented, reusable DAX measures
→ Decision-oriented dashboard design
→ Architecture recommendations

Applied Machine Learning

Predictive models that reach production, not ones that stay in a notebook. Classification, regression, anomaly detection, and clustering on your real data.

Who it's for

Companies that want to predict something specific (churn, demand, failures). Teams with historical data but no model in production.

Includes

→ Feasibility analysis (do your data have enough signal?)
→ Selection and training of the right model
→ Honest results evaluation
→ Clean, documented code ready to integrate

Does not include: deep learning, computer vision, or complex NLP.

Data Engineering

Pipelines that don't break, data that arrives clean and on time, architectures that scale without being rewritten every year. ETL/ELT, data quality, and dimensional modeling.

Who it's for

Companies whose data arrives late, dirty, or not at all. Teams with fragile pipelines. Projects that need an architecture from scratch.

Includes

→ Current architecture diagnosis
→ ETL/ELT pipeline design and implementation
→ Data quality validations
→ Technical documentation and stack recommendations

AI Integrations

Adding AI to your processes in a useful way, without overdoing it. LLM automation, agents, RAG over your documents, and connection to your existing systems.

Who it's for

Companies wanting to automate repetitive tasks with AI. Teams that want to query internal documentation in natural language.

Includes

→ Identification of real use cases
→ LLM integration into existing workflows (Claude, GPT, Llama)
→ RAG over internal knowledge bases
→ Documented, maintainable code

Does not include: model fine-tuning or own GPU infrastructure.

Consulting & Technical Review

1:1 sessions to resolve technical doubts, review architectures, validate stack decisions, or research whether a technology fits your case.

Who it's for

Analysts or freelancers working solo who need a second opinion. Teams wanting to review their architecture from an outside perspective.

Includes

→ Video call session (1h)
→ Code, architecture, or data model review
→ Concrete, documented recommendations
→ Written follow-up after the session

Data Science & Analysis

Exploratory analysis, applied statistics, and segmentation to answer business questions with data. For when you have the data but don't know what it's telling you.

Who it's for

Companies without a data scientist that need ad-hoc analysis. Teams with historical data but no actionable conclusions.

Includes

→ Exploratory data analysis
→ Statistics applied to the business context
→ Segmentation and pattern detection
→ Report with actionable conclusions

Does not include: predictive models (that's Machine Learning, a separate service).

How I work

Tell me the problem

Write to me about what you have, what you need, and what you've already tried. No long forms, no commitment.

Honest diagnosis

I'll tell you if I can help, how I'd approach it, and what it would involve. If you don't need consulting, I'll say so.

We work

Clear agreement, concrete deliverables, direct communication.

Why work with me

I work directly with whoever hires me. No middlemen, no approval chains — what we agree on is what gets delivered.

My stack is what I use daily: Power BI, Python, SQL, Azure, and the LLMs I apply in real projects. What I teach on the blog comes from what actually works in practice.

If I find that your problem has a simpler solution than it seems, I'll tell you. I prefer clients who come back.

Learn more about me →

Got a data project?

Tell me what you have, what you need, and what you've already tried. No commitment — I'll get back to you with an honest diagnosis.