Skip to content
Martin Habedank
Martin Habedank

Head of Product · Agentic & LLM Products · Platform & Data Products in B2B SaaS

Martin Habedank

Berlin, Germany

I lead product in B2B SaaS: I set the strategy, build the cross-functional teams that deliver it, and make them AI-native — because that is where the full potential sits. I also ship products hands-on, which is how I know what that shift actually asks of a team.

Looking for
Head of Product · Director of Product · VP Product in digital & AI-native B2B SaaS — platform & data products. Berlin or remote, open to a fractional CPO mandate.
contact@miskatonic-analytics.com

Hands-on AI work

Products I've built hands-on

Products built end to end under my own label, Miskatonic Analytics, and an open collection of machine-learning models. Not a substitute for a team — the reason I can build teams that work AI-native, and know what that actually demands of them.

  • Miskatonic Analytics

    Dagon — Product-Discovery Engine

    Problem
    Product strategy too often runs on opinion and gut feel. Turning thousands of scattered market signals into evidence-based priorities is slow, manual, and rarely done with rigor.
    Role
    Creator and product lead under Miskatonic Analytics, the product label I founded. I defined the decisions it had to support and built it end to end on an AI-native stack.
    Outcome
    Turns thousands of unstructured market signals into ranked, evidence-backed product opportunities — so roadmap and feature bets rest on data instead of opinion.
    Scope
    End-to-end discovery: from raw market signals across customer personas and segments to prioritized opportunities and a decision-ready view.
    • Aggregates 2,000+ deduplicated findings into Jobs/Pains/Gains per customer persona.
    • Ranks opportunities and maps them by Importance × Satisfaction for clear, defensible prioritization.
    • An agentic LLM pipeline end to end — the kind of build I now expect a team to be equipped for.
    • Agentic & LLM
    • Product discovery
    • Opportunity scoring
    • AI-native delivery
  • Miskatonic Analytics

    RITE — Racing Intelligence & Telemetry Engine

    Problem
    GT racing teams run race weekends on Excel and WhatsApp screenshots. Setup, run, and lap data is scattered, lost between sessions, and almost impossible to compare — so hard-won learning evaporates.
    Role
    Creator and product lead, also under Miskatonic Analytics. I own the problem, the product definition and the roadmap, and built it end to end on an AI-native stack.
    Outcome
    One structured system where engineers and drivers capture and compare an entire race weekend instead of scattered spreadsheets — with AI-assisted coaching insights as the next step.
    Scope
    From problem discovery to a shipped product covering the full race-weekend flow — events, sessions, runs, and lap/sector analysis — aimed at professional GT teams.
    • The whole race weekend in one place — events, sessions, cars, drivers, runs, laps — instead of scattered spreadsheets and chat screenshots.
    • Per-lap and per-sector deltas so teams can actually compare and improve, not just store data.
    • Hands-on proof that a product leader can stay fluent in an AI-native stack — and lead teams that work in one.
    • Data platform
    • B2B SaaS
    • Product strategy
    • AI-native delivery
  • Independent project

    Mobility Model Zoo

    Problem
    Machine-learning problems in mobility are very different from one another: some work with text, others have to run on a microcontroller in the vehicle.
    Role
    Creator. Mobility Model Zoo is my open collection of machine-learning models for mobility.
    Outcome
    Each model solves one concrete problem, is only as large as the job needs, and runs on your own hardware.
    Scope
    One model per problem, from text tasks to microcontrollers in the vehicle. The first models are published on Hugging Face, and many more are planned.
    • Every model card says what the model achieves, measured against what, how fast it runs, and where it fails.
    • No model is released before its licence, privacy and EU AI Act checks pass.
    • Machine learning
    • On-device ML
    • Model cards
    • EU AI Act

    Mobility Model Zoo project page

Who I am

About

I am a product leader who works across strategy and delivery — setting product direction, reshaping how teams ship, and staying close enough to delivery to make sure it lands. Twice a member of the leadership team reporting directly to the CEO.

At CarTelSol I turned a project-based delivery shop into a scalable B2B platform organization: restructuring into cross-functional teams, cutting hardware iteration from over a year to under two months, and moving from years without releases to continuous delivery every two weeks. The platforms my teams shipped serve a global automotive OEM and large fleet operators — from low-code data-science tooling for engineers to an internal AI incubator whose LLM knowledge platform cut compliance research from a sprint to hours.

Under my own label, Miskatonic Analytics, I build products hands-on — Dagon, a product-discovery engine, and RITE, a race-weekend data platform for GT racing teams. That is not an alternative to leading a team; it is what keeps me fluent enough to build good ones. Teams reach their full potential when they work AI-native, and I would rather have done that work myself than ask it of people from the outside.

German
(native)
English
(fluent)

How I lead

Leadership

I build teams and give them clear outcomes and the autonomy to reach them. I am comfortable driving change through an organization as well as through the product itself.

  • Building AI-native teams

    My job is to assemble the team, not to be it. What has changed is what a good team looks like: AI-native ways of working are where the full potential sits, and a team only gets there if the change is real rather than announced. I have restructured a component-owned organization into cross-functional teams with dedicated ownership, and I stay hands-on in the tooling myself — so what I ask of a team is something I have done, not something I have read about.

  • Management by objectives & situational leadership

    I set clear objectives and give teams the autonomy to meet them, adapting my involvement to each team and individual — more direction where it helps, more space where it does not. I rolled out OKRs across engineering to make objectives explicit and shared.

  • Change management

    I led a roughly six-month transformation of a project-based delivery org into a scalable platform organization, navigating resistance by demonstrating measurable velocity gains rather than mandating process. Faster, visible results made the change self-reinforcing.

  • Strategic prioritization

    I invest in platform before features. Putting the reusable foundation first — shared ingestion, tooling, and infrastructure — is what cut hardware iteration cycles from over twelve months to under two and turned multi-year release droughts into continuous delivery.

  • Stakeholder management

    I work effectively across enterprise engineering teams and middle management at a global OEM customer, aligning technical and business stakeholders around shared outcomes and translating between them so decisions actually stick.

  • Company as Code

    Engineering organizations drift without shared standards, repeatable scaffolding, and enforceable quality gates — quality and velocity become person-dependent. I encode the operating model instead: org-wide engineering standards, templates, and automated quality gates baked into the delivery workflow, so good practice is the default rather than tribal knowledge.

Track record

Experience

  1. Jul 2025 – present

    Head of Product · CarTelSol GmbH

    Member of the leadership team (1 of 4), reporting to the CEO/owner — 3 direct reports, ~20 across product & engineering.

    • Redesigned a component-owned organization into 4 cross-functional product teams (Hardware, Firmware, Application, UX), each led by a dedicated Product Owner — owning the operating model end to end.
    • Drove a 6-month transformation that cut hardware iteration from 12+ months to under 2, and release cadence from 3+ years without releases to continuous delivery every 2 weeks.
    • Won executive and engineering buy-in for foundational platform investment over short-term features.
    • Introduced OKRs across the engineering organization, measurably improving alignment and team satisfaction.
    • Platform org
    • Change management
    • OKRs
    • B2B SaaS

  2. Jun 2022 – Jul 2025

    Product Owner / Product Manager · CarTelSol GmbH

    Owned product strategy for the data and telemetry platforms behind an enterprise customer base — Volkswagen and large fleet operators.

    • Delivered a fleet-utilization platform tracking 8,000+ test vehicles (a multi-billion-€ asset pool), enabling a ~50% reduction in fleet investment by surfacing underutilized vehicles.
    • Set strategy and shipped a low-code data-science platform letting engineers analyze vehicle data without programming, cutting iteration from 2 months to under 1 week.
    • Secured government innovation funding off a remote ECU-flashing proof of concept, establishing a recurring innovation-funding function.
    • Co-founded an internal AI incubator; its LLM-powered knowledge platform cut compliance research from ~1 sprint to hours.
    • Data platforms
    • LLM products
    • Enterprise B2B
    • Low-code tooling

  3. 2019–2021

    Chief Marketing Officer · Spiele-Palast GmbH

    Leadership team, reporting to the CEO — owned a €600k marketing budget, built & led a team of 4, drove ~60% of €5–6M annual revenue.

    • Owned marketing P&L and channel strategy across a portfolio of mobile titles; hired and built the marketing team from scratch.
    • Cut marketing-investment payback from ~18 to 6–8 months through cohort and attribution modeling — enabling fast, confident go/no-go decisions on spend.
    • Built the A/B-testing and forecasting engine behind sustained organic and paid growth and seasonal revenue planning.
    • P&L ownership
    • Growth
    • Attribution
    • Team building

  4. 2015–2018

    Marketing Analyst · Spiele-Palast GmbH

    • Established the company's marketing-analytics practice and drove App Store Optimization that measurably grew organic downloads — the foundation that earned the CMO mandate.
    • Analytics
    • ASO

  5. 2012–2015

    Performance Engineer (Telemetry) & Race Operations · Mücke Motorsport GmbH

    • Ran race telemetry and driver performance analysis, feeding real-time insight back to drivers during sessions.
    • Owned race-weekend operations — logistics, scheduling, parts/cost settlement, and warehouse digitization that delivered seven-figure cost savings.
    • Motorsport
    • Data analysis
    • Operations

Education

  • B.Sc. Data Science & Business Analytics · Digital Business University of Applied Sciences · 2021 – present · in progress alongside work · grade average 1.2 (sehr gut)
  • B.Sc. Transportation Systems · TU Berlin · 2008–2013

What I bring

Skills

Leadership & Management

  • Organizational design & scaling
  • Product strategy & roadmapping
  • Change & transformation management
  • OKRs / Management by Objectives
  • Executive & enterprise stakeholder management
  • Data-driven / KPI-led decision-making

Platforms & Data

  • LLMs & agentic / RAG systems
  • Cloud platforms (AWS, Google Cloud)
  • Event-driven & streaming (Kafka / MQTT)
  • Time-series & vector data stores
  • APIs & microservices

Domains

  • Agentic & LLM products
  • Platform & data products
  • B2B SaaS
  • API platforms
  • IoT / Telemetry
  • Automotive
  • Fleet management

Beyond work

Community

  • Co-Founder and Co-Chair, Berlin Chapter

    2024–present

    German Data Science Society

    Co-founded and co-lead the Berlin chapter — building the local data-science community through events and programming.

  • Co-Maintainer

    2024–2025

    Ludwig (open-source deep-learning framework)

    Co-maintained an open-source declarative deep-learning framework, supporting contributors and the community.

  • Co-Founder

    2013–2019

    „Kinderträume" e.V. (German-Russian student initiative)

    Co-founded a student initiative supporting children.

Contact

Let's talk

The fastest way to reach me is email — happy to talk through any of the above in more depth.

contact@miskatonic-analytics.comLinkedInGitHub