Independent enterprise AI advisory

Your AI expert. On demand.

I help technology leaders decide which AI projects to fund, what to build or buy, and how to get promising systems into production.

Working in AI since 2018, with an MSc in Artificial Intelligence and enterprise experience in banking, telecom, manufacturing, and energy.

Dominik Palencar

Meet the founder

Dominik Palencar, MSc

Founder & Principal AI Architect · Prague

Explore my experience

Practical AI experience from real enterprise projects

  • Allianz
  • Raiffeisen Bank International
  • AVL
  • ableneo
  • Orange
  • Slovenská sporiteľňa
  • Česká spořitelna
  • Siemens
  • European Union
  • ABB
  • VSD
  • windCORES
  • TITANS Freelancers

Flagship service · On-demand AI advisory

Independent AI architecture review

Senior AI expertise without the full-time hire.

Bring a decision about an AI system you’re planning, evaluating, or preparing to launch. I’ll review your architecture and options, then give your team a written recommendation covering trade-offs, risks, and next steps.

For teams making these decisions regularly, I also offer ongoing advisory.

Deliverables for On-demand AI advisory
  • Prioritized AI opportunities and a practical roadmap
  • Architecture, model, and vendor recommendations
  • Written trade-offs, risks, and next steps for your team
Strategy
Prioritize the opportunities that fit your goals, data, and team.
Architecture
Review system designs, integration choices, and governance risks.
Emerging AI technologies
Assess where frontier models and agentic systems could be useful.
Engineering decisions
Compare vendors and tools, set evaluation criteria, and give your engineers clear next steps.

Specialist services

AI sovereignty and governance assessment

Assess where your data and models run, who controls them, and which vendors you depend on. Get a prioritized set of governance and infrastructure decisions.

Deliverables for AI sovereignty and governance assessment
  • Inventory of data flows, model hosting, access, and vendor dependencies
  • Comparison of managed cloud, private cloud, and on-premises options
  • Governance gaps and a prioritized action plan

Production readiness review for AI-built applications

Get an independent review of an application built with AI coding tools. I assess its architecture, security, tests, and deployment readiness and recommend what to fix first.

Deliverables for Production readiness review for AI-built applications
  • Architecture, code quality, and maintainability review
  • Security, test coverage, and deployment-readiness findings
  • Prioritized fixes and a checklist for release

Custom frameworks for AI-assisted coding

I set up a project-specific framework for Claude, Codex, and Copilot, including shared instructions, repeatable workflows, and testing and review checks.

Deliverables for Custom frameworks for AI-assisted coding
  • Repository instructions and configuration for your chosen coding tools
  • Repeatable workflows for planning, coding, testing, and review
  • Quality checks, usage documentation, and team handover

Custom AI applications and agentic systems

I design and build AI applications and agentic workflows tailored to your business. We agree on what the system should achieve, how it integrates with your tools, and how to evaluate it before handover.

Deliverables for Custom AI applications and agentic systems
  • Agreed scope, system architecture, and acceptance criteria
  • Custom application or agentic workflow with the agreed integrations
  • Evaluation results, deployment documentation, and handover

AI upskilling and workshops

Build practical AI skills through hands-on workshops tailored to your team’s roles and everyday work. Your team learns to choose suitable tools, use them safely, and check the results.

Deliverables for AI upskilling and workshops
  • Skills assessment and role-specific learning plan
  • Facilitated workshops with exercises based on your team’s workflows
  • Reference materials for tool selection, safe use, and checking AI outputs

AI use case discovery and process assessment

I assess how your business processes work today, including manual steps, handoffs, and recurring decisions. We identify where AI could help and which use cases are worth testing first.

Deliverables for AI use case discovery and process assessment
  • Workflow map showing bottlenecks, decision points, and available data
  • AI use cases ranked by business value, feasibility, data readiness, and risk
  • Pilot brief with scope, an owner, success criteria, and next steps
Orange · ableneo · Designing systems for a 2030 AI strategy

Technical AI strategy and solution architecture

Technical architecture for 22+ telecommunications AI use cases supporting Orange’s strategy through 2030.

The problem

Orange had identified AI use cases for its strategy through 2030. It needed an AI expert to turn those business requirements into technical designs that engineering teams could work from.

My role

As a technical AI consultant through ableneo, I defined solution architectures and advised on technology choices for the identified use cases.

What I did

I reviewed the requirements for each use case, recommended suitable AI models and technical frameworks, and chose how systems should interact. That included deciding when to use event-driven architecture and when a REST API was the better fit.

What I delivered

I delivered detailed solution designs and both high-level and detailed architecture diagrams for 22+ AI use cases across Orange’s telecommunications business.

Orange Slovakia · ableneo · Building reusable skills for AI coding tools

AI-assisted software engineering

A reusable skills framework for Claude Code and GitHub Copilot, designed and implemented for IT professionals across Orange Slovakia.

The problem

Orange Slovakia needed a consistent way for IT professionals to use Claude Code and GitHub Copilot across projects. Teams needed shared practices they could apply in their own workflows.

My role

As a technical AI consultant through ableneo, I was responsible for designing and implementing the skills framework.

What I did

I turned AI-assisted development best practices into shared instructions and reusable skills. I organised them so teams could apply the same practices across projects in both tools.

What I delivered

I delivered a working skills framework for Claude Code and GitHub Copilot, built for all IT professionals at Orange Slovakia.

IPCEI-CIS · TITANS · Designing sovereign AI for industrial energy

Sovereign AI architecture

An open-source agentic system for energy optimisation across industrial and cloud-edge infrastructure.

The problem

Industrial operators need to identify energy-saving opportunities across mixed equipment while keeping control of sensitive infrastructure data and model execution.

My role

As AI Architect on IPCEI-CIS through TITANS, I design and implement agent workflows using open-weight models deployed on premises and MCP connectors to approved knowledge sources.

What was delivered

The system combines internal optimisation playbooks and external technical references to generate recommendations for individual components and the way they are connected.

Why it mattered

On-premises execution supports sovereignty requirements. Evaluating components and their connections helps operators assess energy trade-offs across a site.

Erste Group · Getting enterprise AI ready for banking

AI evaluation and governance

Retrieval evaluation and agentic workflows for Česká spořitelna in a regulated banking environment.

The problem

Early answers from internal documents varied in quality. The team needed a reliable way to measure retrieval and answer accuracy before deploying more AI workflows.

My role

As a founding member of the AI Expertise Team at Česká spořitelna, I co-designed enterprise RAG chatbots and their evaluation pipeline, then engineered components of a production agent platform.

What was delivered

The evaluation pipeline combined automated metrics, qualitative scoring, and failure analysis. The later platform used Microsoft AutoGen to automate text and image extraction across business units.

Why it mattered

In project evaluations, answer quality rose from about 70% to about 95%, supporting production deployment. Reusable components helped teams apply the approach to other workflows.

Orange · Helping sales teams prioritise leads

Applied AI and model selection

A predictive lead-scoring platform to help telecom sales teams decide which prospects to contact first.

The problem

Sales teams needed a consistent way to prioritise prospects by estimated conversion likelihood.

My role

I led the platform development, combining fine-tuned transformer models with a dilated LSTM and a one-dimensional CNN trained from scratch.

What was delivered

The event-driven scoring platform produces a ranking that sales teams can use to decide which leads to review first.

Why it mattered

The scores give teams a consistent input for allocating sales attention to leads with higher estimated conversion potential.

IPCEI-CIS · TITANS · Taking European CO₂ forecasts into operation

Forecasting and MLOps

An open-source data and forecasting platform spanning 43 European ENTSO-E zones.

The problem

A forecasting service across European electricity zones needed reliable data ingestion and a repeatable way to train, evaluate, deploy, and monitor models.

My role

As Senior AI Engineer through TITANS, I designed ingestion and processing pipelines, developed CO₂ forecasting models, and built containerised microservices and MLOps pipelines.

What was delivered

The platform connects model training and evaluation to versioning, deployment, serving, and monitoring across 43 zones.

Why it mattered

A repeatable model lifecycle gives operators a way to track which model is running and monitor its behaviour as forecasts are served.

Siemens · Verifying PLC assembly in real time

Industrial computer vision and augmented reality

A computer vision model ran directly on a worker’s augmented reality (AR) glasses and checked whether a programmable logic controller (PLC) was assembled correctly as they worked.

The problem

Siemens needed to verify that a PLC was assembled correctly during the assembly task.

My role

I built the object detection and classification model that identified PLC components for the assembly check.

What was delivered

The model ran in real time directly on the worker’s AR glasses. It checked the assembly as the worker put the PLC together.

Why it mattered

The model made immediate visual quality checks possible during assembly. Model optimisation and post-processing improved accuracy by 20%.

Confidential client · ableneo · AI research for drone detection and classification

Applied machine learning R&D

Research and development evaluating machine learning for detecting and classifying drones from radio-frequency signals.

The problem

The client needed to find out whether machine learning could reliably detect and distinguish drones using radio-frequency signals.

My role

As a senior AI engineer through ableneo, I led the R&D work on drone detection and classification.

What I did

I developed and evaluated models for both tasks, looking at how reliably they detected drones and distinguished between them in testing.

What I delivered

The project produced research prototypes and evaluation findings that gave the team a basis for judging technical feasibility.

BauHolding · Querying a database by voice

Voice AI and enterprise data

A custom voice bot for BauHolding that answers spoken questions using live data from a MySQL database.

The problem

BauHolding wanted staff to ask questions aloud and get current information from its SQL database within the same conversation.

My role

I designed and built the custom voice bot through dAIp Solutions. I was responsible for how the voice interaction and database access fit together.

What I did

I connected the bot to MySQL through the Model Context Protocol (MCP) and used multi-agent orchestration to coordinate the conversation and data retrieval. I designed the voice bot to work with a retrieval-augmented chatbot.

What I delivered

I delivered a voice bot that takes spoken questions, retrieves live data from MySQL, and answers aloud in the same conversation.

Dominik Palencar

Meet the founder

Dominik Palencar, MSc

Founder & Principal AI Architect | dAIp Solutions

Based in Prague

View my full experience on LinkedIn

My AI work spans banking, telecommunications, manufacturing, and energy, with engagements involving Česká spořitelna, Orange, and Siemens. I’ve also worked with major cloud professional services providers across the DACH region (Germany, Switzerland, and Austria).

How I assess new AI technology

I follow new models, research, and agent frameworks to assess which capabilities are ready for enterprise use and where experimentation is still needed.

I bring an engineering and strategy perspective to AI governance and technological sovereignty.

Ongoing technical advisory

For organizations that need ongoing expertise, I also offer retained AI advisory engagements.

I work alongside engineering teams and technology leaders to review designs, evaluate vendors, and resolve technical decisions as they arise.

Dominik Palencar
Founder & Principal AI Architect
dAIp Solutions

Writing and field notes

See more posts on LinkedIn

More writing

Contact us

Tell me what you’re exploring, building, or trying to improve.

I’ll review your inquiry and, if it’s a fit, suggest a short call to discuss scope and next steps.

You can also message me on LinkedIn.

Project inquiry
Add project details (optional)