Machine learning & optimisation

Prediction and decision systems for complex operations.

I use machine learning and optimisation to turn operational data, objectives, and constraints into models and systems that teams can use with confidence.

Let's build something About me

01 / Profile

JUSTUS KILIAN WOLFF
HAMBURG / GERMANY

Built around the problem, not the method.

I build machine-learning systems, mathematical optimisation models, and the software that makes them useful in practice. The right approach depends on the problem: uncertainty, constraints, trade-offs, or some combination of them.

Today I am building bools.io, a platform for transparent mathematical optimisation. Previously, I developed production machine-learning and optimisation systems at audit.neo. My master’s thesis at Lufthansa Systems focused on aviation planning, and my academic research includes optimisation under uncertainty.

01 Founder of bools.io
02 Former AI Engineer at audit.neo
03 MSc Operational Research with Data Science
04 Published optimisation research

02 / Capabilities

Some problems are uncertain. Others are constrained. Many are both.

Machine learning turns operational data into estimates, signals, and scenarios. Optimisation turns objectives and constraints into plans and recommendations. The right approach depends on the problem.

Understand what is likely to happen.

Machine learning identifies patterns in operational data and turns them into useful estimates, scenarios, and early warnings.

DemandPricesRisk scoresFailure probabilities

Example questions

How many orders will arrive next week?
How are prices likely to change?
Which transactions are most likely fraudulent?
When is equipment likely to fail?

Choose the best action within real-world constraints.

Mathematical optimisation compares possible choices and finds the best feasible option for a defined goal. It accounts for real-world limits such as time, capacity, cost, and availability.

Delivery routesStaff schedulesResource allocationsFacility locations

Example questions

How should shifts be scheduled to meet demand?
Where should a new facility be located?
How should limited resources be allocated?
Which production plan best balances cost and capacity?

03 / Selected work

Built across products, industry, and research.

Decision systems

Making optimisation accessible

bools.io

Building a platform that turns plain-language requirements into transparent mathematical models and solver-backed decisions.

Visit bools.io

Production machine learning

AI tools for audit workflows

audit.neo

Designed and deployed deep-learning fraud detection, an auditor-focused chatbot, and internal machine-learning and optimisation tools.

Visit audit.neo

Mathematical optimisation

Planning aircraft assignments

Lufthansa Systems

Applied problem-specific graph algorithms and column generation to the aircraft tail-assignment problem.

Read about the research

Published research

Locating EV chargers under uncertainty

University of Edinburgh

Co-developed and implemented a stochastic facility-location model. The research was published in Computers & Operations Research.

View publication

View all work

04 / Contact

AVAILABLE FOR SELECTED
OPTIMISATION + ML PROJECTS

Have a prediction or decision system worth building?

Tell me what you are trying to forecast, plan, or improve. You do not need to know whether machine learning or optimisation is the right approach, because we'll figure it out together :)

hi@justuskilianwolff.com