Jacek Janczura Founder-CTO · Enterprise AI
Book a call

Technology leadership · Enterprise AI

From customer boardroom to production code.

I co-founded Jaden Data and led it as CTO to €1M+ annual revenue with no external capital, grew engineering from 0 to 10, and owned AI delivery for APG (a €500B+ pension fund), Bosch, KfW Bank and Boehringer Ingelheim, one of the world's largest pharmaceutical companies — with earlier real-time systems work for ING through SAS Institute. I work from executive discovery and technical pre-sales through architecture, code, production rollout and handover, and I still write the code.

Discuss a leadership role Download the CV
Jacek Janczura

What I have owned

Four things a technology leader is accountable for

Each one with the strongest single piece of evidence I have, and the case behind it.

  1. Customer & commercial

    Discovery, executive workshops, technical pre-sales, architecture and security reviews, scoping and business cases. I am in the room when a buyer decides whether the technology can be trusted.

    Four procurement workflows at a global automotive group taken from a scoped use case to production, with the alignment cadence raised from bi-weekly to weekly and split into parallel sub-groups so no single dependency stalled delivery.

    See the case

  2. Architecture & delivery

    AI systems, enterprise integration, evaluation, observability, reliability, and the rollout into somebody else's production environment.

    A member-facing pension assistant prototyped in a day from platform components and delivered as an on-premises MVP in twelve sprints under Dutch financial regulation; a Dutch ministry highlighted it as one of the best examples of impactful AI integration.

    See the case

  3. Team & operating model

    Hiring, mentoring, delivery standards, roadmap, technical budget, escalation and handover.

    Engineering grown from 0 to 10 inside a bootstrapped €1M+ company, with code-review standards and CI/CD quality gates in place before the team grew, and ISO 27001 and SOC 2 Type 2 delivered in about three months without freezing the roadmap.

    See the case

  4. Reuse & product feedback

    Every deployment leaves something behind — an integration, an evaluation harness, a playbook, a change to the product — so the next one starts further along than the last.

    One specification template and one evaluation template carried a seven-part proof-of-concept portfolio at a German development bank; the process patterns, the cited-retrieval design and the Testbed evaluation framework were reused across later deployments.

    See the case
One engagement becomes the component that makes the next one cheaper.

Nobody handed me a platform, a team or a client list

  1. A company

    Jaden Data was bootstrapped: no external capital, every platform investment paid for by client delivery. It passed €1M in annual revenue while I was its CTO.

  2. A product

    entAIngine grew out of client workflows instead of a speculative roadmap — one multi-tenant, provider-independent platform serving 50+ organisations at 99.9% uptime, rather than a bespoke stack per client.

  3. A team

    Engineering went from my own commits to ten people across frontend, backend and AI workstreams, with structured code review, pair programming and internal knowledge sharing.

  4. A delivery process

    Code-review standards, CI/CD quality gates and testing requirements were in place before the headcount arrived, so production quality never depended on one person being awake.

  5. Client relationships

    Technical pre-sales, project kick-offs, weekly stakeholder cadences and executive reporting were mine to run, for clients in procurement, pensions, development finance and pharma.

  6. A security posture

    ISO 27001 and SOC 2 Type 2 ran as an engineering programme — controls, evidence collection and process changes alongside delivery — and were certified in about three months.

  7. Reusable foundations

    The Testbed evaluation framework, Prompt Wizard and the reusable process components came out of repeated client needs; those components are what turned a one-day prototype into a product.

How I lead

Four rules I do not bend

  1. Understand the workflow and its baseline before choosing technology

    Model choice is the last decision, not the first. What does this work cost today, who does it, where does it break, and what would make the result trustworthy to the person who signs it off?

    In practice At a German development bank every proof of concept started from one specification template — background, problem, objective, scope, methodology, success metrics, dependencies — so seven very different workflows could be compared on the same terms.

  2. Ship the smallest real production loop

    A demo proves nothing about real data, real integrations and real governance. I build the thinnest end-to-end path a real user can run, then widen it.

    In practice APG's assistant was a working prototype in one day from platform components; the twelve sprints that followed went into integration, user validation and the on-premises environment where it had to live.

  3. Treat evaluation, security and observability as product capabilities

    If a system cannot be measured, audited and watched, it is not finished. Those parts belong in the architecture from the start, not after the pilot goes well.

    In practice Testbed made AI processes and retrieval pipelines testable with deterministic cases; ISO 27001 and SOC 2 Type 2 were delivered as an engineering programme in about three months while client delivery continued.

  4. Leave a stronger team and a system somebody else can run

    The measure of the work is what still functions after I stop attending the standup: standards, written decisions, and people who make the calls.

    In practice The code-review rules, CI/CD quality gates and testing requirements outlasted my time as CTO, and the reusable components and evaluation templates are what make each new deployment cheaper than the last.

Where the experience comes from

  1. 2021 – June 2024 as CTO; consulting partner since

    Co-founder & CTO Jaden Data

    Bootstrapped to €1M+ annual revenue, engineering from 0 to 10, 50+ organisations on one platform at 99.9% uptime, ISO 27001 and SOC 2 Type 2 in about three months.

    Owned technology strategy, product roadmap, platform and client architecture, security and compliance, the technical budget, and the engineering organisation from the first hire onwards.

    Read the case

  2. 2020 – 2021

    Software Engineer idealo

    Key contributor to the six-country login rollout that created 1.4M accounts in three months, 700% above forecast.

    Real-time services on Kafka, NoSQL and WebSockets; production microservices in Java and Kotlin; on-call for a platform used by millions of people a month; part of the migration from on-premises to AWS.

    Employer reference (PDF)

  3. 2019 – 2020

    Blockchain Engineer Deutsche Telekom Innovation Laboratories (T-Labs)

    Designed and deployed a Hyperledger Fabric package-tracking network with a secured REST API and integration paths for mobile apps and ERP systems.

    System architecture, chaincode, containerised deployment of the network and its services, and an authentication service spanning blockchain, app and API.

  4. 2016 – 2018

    Associate Technical Consultant SAS Institute, on assignment with ING

    Real-time behavioural decisioning at thousands of events per second — one of the first SAS Event Stream Processing implementations in the European Union.

    Data pipelines on Kafka, Flink and Camel, and a GDPR data-lineage plugin for the ETL tooling that was adopted by analysts across ING Group.

Available for full-time leadership roles

Remote-first, on-site when it matters · Polish (native), English (fluent), German (conversational)

I did not stop writing code when I started leading

  1. Taufolio

    A current build of my own: an AI equity-research system that turns SEC filings, earnings-call transcripts and market data into cited reports. Parallel multi-agent drafting with reconciliation and citation anchoring, so every material claim links back to the source document it came from.

    taufolio.com

  2. RP-Matcher decisions

    For an emergency-lighting manufacturer I engineered the features, benchmarked five model families with hyperparameter search, and chose the tuned random forest, which reaches 78.4% top-1 accuracy on the current corpus — then wrote down what more labelled pairs would buy, instead of reaching for a deep model the dataset could not support.

  3. Code and tooling in the open

    Repositories, experiments and the tooling I use day to day are on GitHub.

    github.com/jjanczur

  4. Talks and publications

    Conference talks on blockchain research and applications, and peer-reviewed publications on decentralised compensation of digital content, trusted inter-organisational data processing and static vulnerability analysis.

    Talks and publications

Let's talk about the role

Tell me about the team, the product and what has to be true in the first six months. I will tell you straight whether I am the right person for it.

Discuss a leadership role Read the case studies

Available for full-time leadership roles · remote-first, on-site when it matters