Senior Scientific Software Engineer — Simulation and Machine Learning
LocationConstructor Labs, Germany
Work modehybrid
Typefull-time
SenioritySenior
DepartmentResearch
Company size201–500 people
PostedSep 21, 2026 · 2w ago
Verified live8h ago
At a glanceSummarised by Seekless from the posting.
Must have9
PhD or equivalent in physics, applied mathematics or computational science
Strong scientific computing in Julia, Python, JAX or comparable
Real numerical depth: ODE solvers, optimisation, stochastic processes, automatic differentiation
Published work, or code that others rely on
Able to take a method from a paper and make it fast, tested and usable
Comfortable owning code quality, not only correctness
Interest in supervising students alongside the engineering
Clear technical writing
Fluent English
Skills
Julia
Python
JAX
Julia core
Python interface
ODE solvers
automatic differentiation
neural-network approaches to PDE solving
equation discovery
establish academic tools
benchmark
simulation code
differentiable simulation
model learning
parameter inference
optimisation
numerical simulation
machine learning
About the role
Location: Remote within Europe, with regular presence in Bremen
Direct Reports: None; supervision of students and interns
Working Model: Remote / hybrid. Regular time on-site in Bremen.
Mission
Build the simulation and model-learning engine at the centre of Constructor’s quantum software. The work is numerics-heavy: differentiable simulation of physical systems, inference of model parameters from experimental data, and optimisation at scale. The role sits between physics and engineering and requires genuine depth in both.
Responsibilities
Pillar 1 Simulation Engine
•
Build and maintain high-performance simulation code for open and closed quantum systems.
•
Own the differentiable simulation and automatic differentiation layer.
•
Drive performance. This code runs inside optimisation loops and speed decides what is possible.
•
Maintain and extend the Julia core and its Python interface.
Success Metrics and KPIs: simulation accuracy against measurement, runtime performance, API stability.
Pillar 2 Model Learning and Optimisation
•
Implement parameter inference and model learning from experimental data.
•
Work on neural-network approaches to PDE solving and equation discovery.
•
Benchmark against established academic tools and publish the comparison.
Success Metrics and KPIs: benchmark results, methods adopted by the research group, publications.
Pillar 3 Research Collaboration
•
Work with Constructor University faculty on joint research and grant applications.
•
Supervise graduate students and interns working on the engine.
•
Contribute to course material on numerical simulation and machine learning for physics.
Accountability: research collaborations initiated and sustained; students productive.
Requirements
Experience
•
PhD or equivalent in physics, applied mathematics or computational science.
•
Strong scientific computing in Julia, Python, JAX or comparable.
•
Real numerical depth: ODE solvers, optimisation, stochastic processes, automatic differentiation.
•
Published work, or code that others rely on.
Skills & Competencies
•
Able to take a method from a paper and make it fast, tested and usable.
•
Comfortable owning code quality, not only correctness.
•
Interest in supervising students alongside the engineering.
•
Clear technical writing.
Education & Languages
•
PhD in physics, applied mathematics or computational science.
•
Fluent English.
What This Role Is (and Is Not)
This role IS
•
A research-grade engineering role with a product at the end of it.
•
Central to the group’s core technical asset.
•
A role with a route into teaching and supervision if you want it.