BARTAS CONSULTING

Selected work

From a strategic question
to something you can inspect.

6 examples across banking, product, research and integration, ordered by project start, newest first. Explore the question, my contribution and the linked product or source.

David Bartas, founder and technology advisor
01 / BuildPortfolio.ai

Keep the investor’s intent visible in every revision.

  1. 01Objectives & rules
  2. 02Editable proposal
  3. 03Compare & refine
Conceptual workflow · Explore the linked product or source

Product · Portfolio construction

BuildPortfolio.ai — case study

Oct 2026 – present

In development

Product direction & architecture · Investboard GmbH

Turning an investor’s objectives into a portfolio they can review and refine.

The question
How do you move from an investment objective to a complete proposal while keeping constraints, holding decisions and evidence visible?
The work
Design a focused construction workspace around the investor’s brief. Keep the portfolio itself central, with direct editing, holding locks and saved revisions, while connecting research and analysis to Investboard.
The product design
  • Objectives, investment rules and separate holding and weight locks
  • Editable proposals, saved variants and comparison of changes
  • Research handoff and an integration with Investboard’s shared services
Current stage
The website preview is public. The workspace is being prepared for shared account access and live AI, with full service acceptance and the buildportfolio.ai domain launch still pending.
Explore BuildPortfolio’s previewAbout Investboard
02 / InvestingStack

Build trust into the content and the data model.

  1. 01Source evidence
  2. 02Structured research
  3. 03Reusable records
Conceptual workflow · Explore the linked product or source

Research · Editorial product

InvestingStack — case study

Sep 2026 – present

Creator & editor

A research product that makes AI systems easier to inspect.

The question
Repository popularity and broad AI claims tell readers little about how a system works, which data it uses or where its limitations sit.
The work
Structure the ecosystem into a browsable directory and publish research with explicit sources, version boundaries and reusable evidence. Make the distinction between a documentation review and an observed test visible.
What was delivered
  • Directory of AI agents, models, data and infrastructure
  • Research articles with linked sources and methodology
  • Downloadable audit data and a reproducible case-study protocol
The public result
The English and German site publishes an audit of 20 open-source repositories and a connector case study. Readers can inspect the supporting data, methods and stated limitations.
Visit InvestingStackRead the ecosystem audit
03 / Integration

Design the integration around a traceable hand-off.

  1. 01Investor context
  2. 02Agent analysis
  3. 03Recorded check
Conceptual workflow · Explore the linked product or source

Open source · Integration

TradingAgents × Investboard — case study

Sep 2026

Project in the Investboard ecosystem

Connecting a local agent run to a reviewable decision record.

The question
A research framework’s output can be separated from the investor’s mandate, current position and the record used to review a decision later.
The work
The connector supplies Investboard-backed data, policy and position context to TradingAgents, then posts the completed analysis back to Investboard for storage and a mandate check.
What was delivered
  • Documented local setup using the user’s own model keys
  • Policy and position injection into agent context
  • Completed-analysis posting and recorded mandate verdict
The public result
The repository provides source, setup instructions and tests. Its documentation names TradingAgents v0.4.0 as the tested release and lists historical-data exceptions. A verdict records a check; order blocking needs enforcement in the execution path.
Inspect the connectorRead the documented case study
04 / Investboard

Make context a part of the product architecture.

  1. 01Your mandate
  2. 02Portfolio context
  3. 03Review & record
Conceptual workflow · Explore the linked product or source

Product · Data · AI context

Investboard — case study

2025 – present

Founder & architect

A portfolio platform built around the investor’s own plan.

The question
Portfolio data is spread across accounts. An AI assistant also needs the investor’s rules and reasoning to put that data in context.
The work
Connect the product experience, portfolio calculations and permissioned agent access around a shared mandate. Keep the investor’s decision history available alongside the current holdings.
What was delivered
  • Portfolio overview and allocation analysis
  • Personal mandate and recorded checks
  • Decision journal and context for supported AI agents
The public result
A public product and interactive demo let visitors explore the workflow. The demo identifies example data; a recorded mandate check leaves the investment decision with the user.
Explore InvestboardView the product demoAbout Investboard
05 / George Business

Connect the customer journey to the system architecture.

  1. 01Business requirements
  2. 02Architecture & APIs
  3. 03Feature delivery
Conceptual workflow · Explore the linked product or source

Banking · Payments · Integration

George Business — case study

Aug 2022 – Jun 2025

Senior Analyst · Core delivery squad

Connecting payment journeys to the systems behind them.

The question
How do you turn a complex payment lifecycle into a clear experience across internet and mobile banking?
The work
My role spanned solution architecture through feature delivery, connecting business requirements, APIs and system integrations within the core squad.
My contribution
  • Payment lifecycle views in internet and mobile banking
  • Transaction history and card reservations
  • API testing and collaboration across analysis, design and delivery
Explore the product
George Business is Česká spořitelna’s corporate digital banking platform. The bank’s public product page shows the wider service; my LinkedIn describes my role and scope.
Role and scope on LinkedInExplore George Business
06 / BankID onboarding

Design the journey and the visibility needed to run it.

  1. 01Identity verification
  2. 02Onboarding integration
  3. 03Process monitoring
Conceptual workflow · Explore the linked product or source

MONETA Money Bank · Digital onboarding

BankID onboarding — case study

Jun 2021 – Feb 2022

Lead Analyst

Making digital identity a reusable part of onboarding.

The question
How do you integrate digital identity across onboarding journeys and make an automated process observable?
The work
I translated business requirements into solution architecture and user stories, coordinated system changes, and worked with a business colleague on the onboarding journey.
My contribution
  • BankID integration across web and mobile onboarding
  • A reusable KYC verification flow
  • KPI reporting and alerts for potential process issues
The project result
The verification flow was rolled out across web and mobile onboarding. My LinkedIn records the project scope; MONETA’s public page explains its bank identity service.
Role and scope on LinkedInBank identity at MONETA

The banking cases describe David’s personal project roles. The other examples cover his own products and work in the Investboard ecosystem. Follow the links to explore the products, inspect the source and read the research.

A clear place to start

Your next decision.
Let’s work it through.

Choose a focused introduction or a longer working session. Every consultation is with David Bartas personally.

  • Monday–Friday, 18:00–21:00 Berlin time
  • Online via Google Meet
  • Book at least 48 hours ahead. Google displays slots in your local time zone.

30 minutes

€100

Introduction

Bring one decision. Leave with a clearer next step.

A focused conversation to frame your AI, product or architecture question, review the constraints and decide where to start.

  • Clarify the problem and priorities
  • Discuss fit, scope and the next step
Book 30 minutes

90 minutes

€250

Strategy session

Work through the options and build a practical plan.

A working session on a defined challenge: compare approaches, identify trade-offs and shape a plan for your product, AI initiative or technology architecture.

  • Map options, dependencies and risks
  • Agree priorities and an actionable plan
Book 90 minutes

Choose a session, select an available time and add a few lines about your question. Google sends the confirmation and meeting details. These are the same booking pages used by InvestingStack.

Have a larger engagement in mind? Email your brief to discuss the scope. Reviews, workshops and delivery support are quoted separately.

A little preparation

Before you book.

Start with the question that matters most. A clear goal helps us make the time useful.

Which session should I choose?

Choose the 30-minute introduction to frame a question, discuss fit and identify the next step. Choose the 90-minute strategy session when you have a defined challenge and want to compare options, dependencies and priorities in more depth.

What should I prepare?

Bring one decision you need to make, the outcome you want and the main constraints: timing, budget, systems or team capacity. Add a short summary in the booking form. A roadmap, system diagram or product brief can help give the conversation context.

How do time zones and availability work?

Sessions are available Monday–Friday, 18:00–21:00 in Europe/Berlin, with at least 48 hours’ notice. Google shows available times in your local time zone and sends the confirmation and Google Meet details. Both websites use the same calendar, so a slot booked through InvestingStack is unavailable here too.

Can we work together beyond a consultation?

Yes. Architecture reviews, product and AI workshops, and delivery support can be scoped separately. Email a short brief with your goal, current situation, timing and the help you need. The next step is to agree the scope and fee.