About FION
European industry is losing competitiveness because electricity here is more expensive and more volatile than in the US or China. The reason: renewables fluctuate strongly, factories consume constantly.
We close that gap. FION plans and installs the right battery system for an industrial site and runs it with AI in real time against tariffs and markets. The software that dimensions, controls and monitors those systems is our own, and it runs end to end: ML and optimization in the cloud, data streaming and control, and the edge device at the customer site. The result for the customer: significantly lower energy costs and measurable CO2 savings. We earn trust by delivering systems and operating them transparently.
We already work with more than 20 factories across industries such as plastics, automotive and food processing, and we are backed by early-stage investors.
We're looking for an engineer with deep experience in time-series forecasting and mathematical optimization to join us as an early hire. You'll work on the forecasting and optimization core of our platform: the models that forecast industrial electricity consumption and PV generation, and the optimization that decides how each battery operates and trades against tariffs and markets within physical and grid constraints. This is a startup role: you should be comfortable switching between strategic thinking and hands-on work.
You'll work directly with the CTO, one of the co-founders.
Tasks
Forecasting models Develop and improve time-series forecasting models for industrial electricity consumption, PV generation, and other energy-relevant signals, and extend them with probabilistic outputs. Evaluate forecast quality on noisy, incomplete, and non-stationary industrial data, including its actual operational and economic impact.
Battery dispatch optimization Develop and improve the optimization models that schedule battery dispatch across peak shaving, self-consumption, spot market trading and flexibility marketing, while adhering to physical and grid constraints. Design how forecasts, uncertainty measures, physical constraints, and system states are used in downstream optimization workflows.
Validation, simulation and monitoring Build and improve simulation, replay, benchmarking, and validation workflows to test model behavior before and alongside deployment in live systems. Build and use tools and processes to monitor and assess the quality of operational forecasting and optimization models.
Production integration Design, write, test, and deploy production-grade code for mission-critical forecasting and optimization products. Improve robustness through plausibility checks, fallback behavior, re-forecasting, and handling of low-confidence or missing data. Collaborate closely with software engineers to integrate models into our core platform and edge devices.
Requirements
You bring:
5+ years of professional software engineering experience
Strong applied experience in time-series and probabilistic forecasting: forecast calibration, uncertainty evaluation, backtesting, and model validation
Strong understanding of mathematical optimization, especially LP/MILP, and the ability to model real-world systems through objectives, constraints, and operational rules
Strong Python skills with the scientific and machine learning tool stack (pandas, NumPy, SciPy, scikit-learn, PyTorch)
Experience developing, releasing, and tracking the performance of forecasting or optimization models in a commercial software setting
An ownership mindset - you want to shape the platform, not just implement tickets
Even better if you have:
Energy domain knowledge - you understand power, energy, phases, and how the grid works
Understanding of European electricity markets, flexibility services, and grid operation
BESS-specific experience: familiarity with BESS architectures, components, and operation, and direct experience with energy management algorithms for battery systems
Experience with open-source and commercial MILP solvers (e.g. Pyomo, OR-Tools)
Experience with optimization techniques such as stochastic or robust optimization
Experience building forecasting and machine learning products in the cloud (e.g. AWS, GCP, Azure)
Experience using Large Language Models to accelerate software development and reliably add new capabilities to real products
German language skills (our customers and partners are German)
Benefits
Ownership, not tickets This isn't a role where the models are decided and you fill in the blanks. The forecasting and dispatch optimization stack is greenfield - nothing exists yet, and you'll design and build it from the ground up. At larger energy companies, this scope is split across multiple teams. Here, you own it all.
Real impact, fast Your code runs physical batteries that save real factories real money. You'll see results on a Grafana dashboard the same week you deploy.
Founding trajectory As FION grows, you grow with it - from hands-on builder to platform lead to head of forecasting and optimization. What takes 5 years at a large company, you'll have from day one.
Compensation
VSOP (virtual shares)
you participate in the company's success
Berlin hybrid (remote from within Germany possible for the right person)
Send us:
Your CV or LinkedIn profile
A few sentences on why this role interests you
Optionally: a link to a project, repo, or writeup that shows how you think about technical problems
We review every application personally and get back to you within a week. No automated screening, no ATS black hole - you're writing to a person, not a system.
