End-to-end product management in IoT, smart home and home energy management (HEMS), now with applied AI.
Up to 18%
Heating energy saved through a SaaS intelligent heating control.
€130k
Awarded EXIST pre-seed funding from the Federal Ministry (BMWi) for a 0-to-1 start-up.
Dual MSc
ENTECH at KIT and IST Lisbon, with an EU InnoEnergy scholarship (€20k).
The journey, zoomed out
From one device in one home to the energy system.
Every role widened the lens. Follow the drawing as it zooms out, and see what each step changed.
Scale 1 mA single device01 / 05
01 · SmartHome Configurator (Smartivate) · 2018 to present
A home full of devices that do not talk.
90+ manufacturers and 6+ conflicting standards. A heat pump or a smart thermostat only saves energy if it works with the rest of the house. I co-founded Smartivate to make sure it does.
450+
devices mapped for compatibility
2,500+
users since launch in 2020
€130k
EXIST funding secured
ImpactFewer incompatible purchases, so more homes can adopt efficient devices with confidence.
Heat is lost where nobody looks. At Uponor I put sensors on the boiler and in the coldest room, so the heating could respond to real conditions, and piloted it in multi-family, mixed and commercial buildings.
12–15%
heating energy optimised in pilots
3
European markets
€0
upfront cost to the customer
ImpactSavings are measured against a baseline, so the customer pays only from energy actually saved.
03 · Intelligent Heating Control (E.ON One) · 2023 to 2026
A building that heats itself smarter.
Intelligent Heating Control is an IoT heating SaaS. It cuts the energy a building uses without asking tenants to change anything. If the box ever loses its connection, it falls back to bypass mode and the heating keeps running.
Up to 18%
less heating energy
400+
field installations
7–10 yr
recurring SaaS model
ImpactLess fuel burned in every connected building, every heating season.
80%+ of heating systems run inefficiently. With Danfoss Leanheat, an AI algorithm optimises district heating, gas and oil boilers, and our joint go-to-market started in Germany and Finland with 5 successful pilot buildings.
80%+
of heating systems inefficient
52 TWh
annual waste, at most
€3B+
wasted a year in Germany
ImpactProducts that scale through partners, so one good idea reaches many buildings.
Heat pumps, home energy management and flexible demand all need software that decides well. After the Ironhack AI Consulting and Integration programme (400+ hours), I build agents and copilots for exactly these problems.
ImpactEnergy products that learn, so efficiency does not stop at installation.
I am dedicated to steering my work towards carbon neutrality and efficiency: products that help buildings use less energy, and energy systems run on less waste, every year from here on.
Co-founder and Product Lead, 2018 to present. I built Germany's first IoT SaaS configurator, designed to remove smart home fragmentation and speed up energy efficiency.
In shortCo-founder and Product Lead of Germany's first cross-brand IoT SaaS configurator: 450+ plug-and-play devices mapped for compatibility, 2,500+ users, EXIST-funded.
Days or weeks of research to find the right device among 600+.
Tailor-made, reliable solutions generated in under 20 minutes.
46% of users fear incompatibility, and return rates are high.
Cross-brand interoperability guaranteed through proprietary mapping.
Isolated devices leave 55.8% of users afraid of privacy and security gaps.
Cohesive ecosystems built for energy efficiency, intelligence and security.
The PCMS engine
The Product Compatibility Mapping System cross-references a home against 450+ plug-and-play devices and returns bundles that are known to work together.
InputHome attributes, infrastructure layout and budget constraints.
MappingPCMS checks every candidate against the full compatibility database.
OutputBest-rated, fully compatible device bundles for heating, lighting and security.
The compatibility database as a rack.
A B2B2C business model
B2B partners: real estate and utilities
Paid a monthly SaaS subscription plus a setup fee.
Status A model I designed and tested in Smartivate's early years. It is no longer part of the business.
Platform and retail partners
Pays a 3 to 5% basket commission and €0.20 per click.
Traction Lower product return rates thanks to guaranteed compatibility.
B2C: homeowners and tenants
Gets personalized consultation and tailor-made packages.
Traction 2,500+ users served since the February 2020 launch.
Funding, testing and growth
Funding. Awarded €130k EXIST pre-seed funding from the Federal Ministry (BMWi), plus four accelerators including AXEL Energy Accelerator.
Pre-launch testing. The configurator went through monthly software testing events at CyberForum Karlsruhe, with 5 to 7 testers each month for 7 to 9 months before launch.
Growth. End-to-end SEO, including configurator links in smart home forums and backlink building. Google Analytics and heatmaps tracked drop-offs, button clicks and regional access across Germany.
Recognition. hubitation finals winner 2018, 3rd place at the SmartHome Deutschland Award 2019 and Proptech Top 100 2020. See the Credentials tab.
1 → N
EcoMate (Uponor)
Product Manager, IoT and Digital Services, July 2021 to July 2023, Frankfurt. I worked on connected building products for three European markets, from sensor pilots in real buildings to a business model that costs the customer nothing up front.
In shortProduct management for sensor-based heating optimisation across Germany, Spain and Denmark: pilots in multi-family, mixed and commercial buildings and a savings-share model with no customer CAPEX.
Optimising heating with sensors
Connected sensors let a building's heating respond to what is really happening inside it, instead of running on fixed settings.
12–15%
heating energy optimisation in the first pilots
3
building types piloted: multi-family, mixed and commercial
3
European markets served
Looking for the loss inside the building.
How the closed loop works
01 · BoilerSensors in the basement, connected to the boiler, measure and optimise heating energy consumption.
02 · Coldest roomA sensor in the coldest room of the building reports back how the heating actually arrives.
03 · FeedbackThat reading closes the loop, so the boiler settings follow real conditions in the building.
04 · ResultHigher energy efficiency of the building, without changing how tenants live.
Pilots across building types
We ran pilots in multi-family houses, mixed-use buildings and commercial buildings. They achieved a heating energy optimisation of 12–15% as a start, which gave us evidence from three very different building types rather than one.
Market analysis and business model
Competitive analysis. I ran a rigorous competitive analysis to understand how the market works, and developed the business models from it.
Pay from the savings. One model measures baseline consumption against actual consumption after the sensors are installed. We earn revenue from the energy saved in each building and take a share of it from the customer.
No CAPEX for the customer. The customer pays nothing up front. The cost is covered from savings that are measured, so the incentives of both sides point the same way.
Delivery and team
Remoni KPIs. Worked with the Remoni development and sensor teams to define the KPIs shown in the web app's frontend, and coordinated with Remoni stakeholders in Denmark.
Customer web app. Defined the roadmap for the development team in Spain to build the web app interface for customers.
Leakage detector calculator. Designed a calculator that quantifies the loss from pipe leaks inside buildings.
Team leadership. Led the Spanish team, based in Spain, as part of the international product setup.
1 → N
Intelligent Heating Control, IHC (E.ON One)
Product Manager, Intelligent Heating Control (IHC), August 2023 to March 2026, Essen. I took start-up speed into corporate scale, aimed at the €3B building energy problem.
In shortProduct Manager for IHC at E.ON One: four business models, a Danfoss partnership, a 200+ installer network and a fail-safe rollout.
Intelligent Heating Control
An IoT-based smart heating SaaS, built with partners including Danfoss Leanheat, Roamler and Lemonbeat.
Up to 18%
reduction in heating energy consumption in the best installations
400+
IHC-Box field installations deployed
7–10 yr
recurring SaaS revenue model built with Danfoss Leanheat
Energy, one connection at a time.
The Danfoss Leanheat partnership
Split of work. E.ON One supplied the IHC boxes. Danfoss Leanheat's AI algorithm optimised heating energy for district heating, gas and oil boilers in Germany.
Go-to-market. A joint go-to-market partnership starting in the German and Finnish markets, with 5 pilot buildings in Germany, all successful.
Business model. Moved from one-time hardware sales to a jointly run offering.
My ownership. Product planning, technical support, the business model and legal coordination on the E.ON side, with support from sales, marketing and operations.
Scaling installations through a partner network
To get IHC into hundreds of buildings, I brought in Roamler, whose nationwide network includes 200+ installers, and trained them to install the box in the basement of multi-family buildings. I then designed the workflow that let four companies deliver installations together.
200+
installers in the partner network, trained on the IHC box
4
companies working as one delivery team
OrchestratorE.ON OneMy role
TechnologyLemonbeatTechnical support
FieldRoamler200+ installers
CustomerReal estateMulti-family buildings
Installers report back to E.ON One
Four organisations, different systems and goals, one closed loop with a clear owner for every handover.
Four business models I designed
I designed and tested four ways to bring Intelligent Heating Control to market, so the choice of model came from evidence rather than assumption.
MODEL 01
Hardware sale plus multi-year SaaS
Direct sale of the hardware with a multi-year SaaS contract. The classic model, and the one most customers understood fastest.
Most easily understood
MODEL 02
Rental
A monthly fee over a multi-year term, with no upfront hardware cost. We offered it as part of A/B testing and no customer chose it. That is useful evidence in itself about how this customer group budgets.
A/B tested
MODEL 03
Partner model with O2 Business
The telecoms provider brought customers to us. This turned their sales channel into our acquisition channel.
Partner channel
MODEL 04
Shared model with Danfoss Leanheat
Our hardware and field operations, their optimisation algorithm, in place of a one-time hardware sale.
Joint offering
Built to fail safely
If the IHC box loses its connection, it switches to bypass mode. The outdoor temperature sensor signal passes straight to the heating system, so tenants keep their heating.
Alongside the Danfoss teams, I worked with E.ON's own engineering team on product development, and navigated EU energy regulation (GEG, EnWG) to launch the enterprise SaaS. The work earned the Best of Technology Award 2024, shown in the Credentials tab.
Core PM toolkit
Domain knowledge, delivery methods and the tools I use every day.
Domain expertise
IoT and smart home mesh networks
Home energy management systems (HEMS)
Distributed energy resources (DER)
Heat pumps and EV chargers
EU energy regulation (GEG, EnWG)
Cloud, APIs and interoperability
Product management
Roadmaps
Product strategy and vision
Product discovery and customer research
Evidence-based prioritisation
Backlog ownership and user stories
KPI definition and product analytics
User research and UX
Hardware and software integration
AI and automation
LLM apps and RAG
Agents with n8n and LangGraph
Evaluation with LangSmith
Streamlit prototypes
Stakeholders and leadership
Stakeholder management
Cross-functional leadership
Distributed, international teams
Strategic partnerships and ecosystems
Startup mentoring
Agile execution
Scrum, Kanban and OKRs
Lean Startup
Design Thinking
Commercial operations
Business case and go-to-market
B2B2C platform dynamics
Recurring SaaS models and pricing
Start-up funding
Product and delivery tools
Jira
Confluence
Miro
Microsoft Office (Excel, PowerPoint, Word)
Analytics and BI
Google Analytics
Heatmaps and session analysis
Business intelligence dashboards
Tableau
Languages
English
German
Telugu
Hindi
Tamil
Credentials
Education, certifications and recognition for products I built and teams I led.
Education and certifications
2026
KI-Manager/-in (IHK)
IHK Akademie München und Oberbayern. Seminar completed with a jury presentation, certificate dated 15 October 2026. View on LinkedIn
2026
AI Consulting & Integration, Certificate of Completion
Ironhack. A 400+ hour program, completed 17 September 2026. Projects are on the AI projects tab. View on LinkedIn
Product Academy, Zurich, Switzerland. Certification completed August to December 2024. View on LinkedIn
Degree
MSc ENTECH
KIT Karlsruhe and IST Lisbon, with an EU InnoEnergy scholarship (€20k).
Programs
Accelerators
Participated in four accelerators, including AXEL Energy Accelerator.
Awards
2024
Best of Technology Award
Awarded to E.ON One for Intelligent Heating Control.
Award ceremony, 2024.
2020
Proptech Top 100
Smartivate.
2019
SmartHome Deutschland Award, 3rd place
Category Studentische Leistung (student achievement), for Smartivate, the configurator for plug-and-play smart home devices.
Berlin, 2019.
2019
Energy Efficiency Hack, GETEC Challenge winner
Winner of the GETEC challenge at the Energy Efficiency Hack 2019, organised by DENEFF.
Energy Efficiency Hack, April 2019.
2019
InnoAlliance Xi'an, 4th place
International innovation competition.
2018
Energy Efficiency Hack, overall winner
Overall winner of the Energy Efficiency Hack 2018 in Berlin, organised by DENEFF.
Energy Efficiency Hack, Berlin, 2018.
2018
hubitation finals, winner
Smartivate won a place at EXPO REAL 2018, the real estate trade fair.
hubitation finals, 2018.
AI projects
Working prototypes from the Ironhack AI Consulting and Integration bootcamp (July to September 2026), and from preparing for product manager interviews. Each one starts from a real problem in energy, heating or small business.
Move across the board to raise a commit.
IRONHACK BOOTCAMP
CapstoneIndividual, September 2026
Heat Pump Copilot
A commissioning and connectivity copilot for installers at small and mid-sized heat pump vendors entering the German market. It tackles two field problems: fault codes that installers cannot resolve on site, and controller connection drops that customers mistake for hardware failure.
Retrieval over manufacturer manuals, a rules-based commissioning completeness checker, an early-warning tool for drops in heat pump efficiency (COP), tracing with LangSmith and an LLM-as-judge evaluation inside the app. Round 1 covered market sizing and a Tableau dashboard. Round 2 covered EU AI Act classification, GDPR, ROI and risk, and deployment strategy.
An n8n workflow that takes school book orders from a family bookshop's Telegram chat. It reads text and photographed handwritten orders, matches them against the publisher catalogue, logs each order and answers status questions. On request it consolidates open orders into one list per publisher, ready to send.
Catalogue extraction from a 44-page publisher price list PDF, a daily dashboard for the shop owner, and an audit against the EU AI Act (minimal risk, with a transparency duty) and GDPR for a possible rollout to other independent booksellers.
n8n
Telegram
OpenAI vision
Google Sheets
Project 2Team project, 2 days
Feature Launch Kit
Turns a feature's documentation into launch material: a marketing email, a user guide and a sales one-pager. The MVP injects the document directly into the prompt and runs two language models, OpenAI as primary and Cohere as fallback.
My part: the LLM client, including the fallback logic, and the sales one-pager prompt template.
Python
OpenAI
Cohere
Streamlit
Agent labsJuly to September 2026
Agents, RAG and automation labs
A LangGraph complaint processor with a conditional validation step, an n8n Telegram-to-Airtable logger, a study of two n8n RAG workflows built on Pinecone, OpenAI embeddings and a Cohere reranker, and a comparison of chunking strategies for a PDF and a podcast recording.
LangGraph
n8n
Pinecone
Airtable
BUILT FOR INTERVIEWS
Product manager, energy marketInterview showcase, 2026
Delivery Lifecycle Diagnostics Copilot
A supplier-switch funnel dashboard with an AI triage agent that explains where and why customer switches stall. It models the lifecycle across Germany, Norway, Sweden and the Netherlands, including each market's own rules: bilateral switching under GPKE in Germany, the central Elhub in Norway, and EDSN in the Netherlands.
An actor and process map, a boundary memo on foundational delivery versus contract logic, a natural-language query layer over the case data, and an integration roadmap for connecting to real systems. Case data is synthetic.
A Streamlit app for AI-supported product portfolio management. An LLM intake step flags weak ideas with a disproof signal, ideas are scored on weighted criteria and move through Explore, Validate, Scale and Stop gates, and each business case is tested for sensitivity. Built around an illustrative data center cooling concept.
Streamlit
LLM intake
Stage gates
Business case model
Ready to bring smart energy products to the next level.
Open to senior product roles in IoT, smart buildings, energy and applied AI. Book a virtual coffee to get to know me, or message me on LinkedIn.
