We build PULSAR, a platform for teams integrating complex cyber-physical products. PULSAR reconstructs how each system version actually behaves, giving engineering teams at OEMs and along the supply chain straight answers.
The Challenge
We spent decades inside automotive and SDV programs at Audi, BMW, CARIAD, Elektrobit, Tasking, and Qualcomm. We saw brilliant engineering teams lose weeks chasing incidents that should have been resolved in hours.
The problem was not a lack of data. It was the gap between seeing data and understanding system behavior. Today’s tools show vast amounts of information. They do not explain how complex systems actually work, interact, and fail.
This is why we founded AIORX.
Alejandro Vukotich · Gregor Zink · Alfons Pfaller · Ewald Gössmann
AIORX Founders · 25–30+ yrs each at Audi BMW CARIAD Qualcomm Ericsson Elektrobit Siemens Tasking Continental + others
The shift
Traditional approaches keep adding complexity.
More documentation, more people, more processes.
Eventually, complexity grows faster than teams can manage it.
Execution slows down. Problems take longer to understand and resolve.
The good news: data can reveal a lot about your system, and it is already available.
What’s missing is turning that data into understanding.
That’s
what PULSAR does.
What we do
With every new software candidate you start testing and recording data. The PULSAR engine uses this data to reconstruct how your system actually behaves, and how every component works. That model becomes your ground truth: explore it yourself, or let an agent surface what matters.
PULSAR modules
view
System Explorer The architecture as builtview + agent
Anomaly Detection Every run against a learned normalview + agent
Change Detection What moved between two buildsagent
Root-Cause Analysis Symptom back to originagent
Agent Builder Your own questions, kept as a skillPULSAR engine
No specifications. No documentation. No source code.
Why PULSAR
01
1,000s
interactions per run — turned into reusable system knowledge
Every release adds components, suppliers and variants — but also new runtime evidence. PULSAR learns from every run, so its understanding of normal behavior, dependencies and failure propagation grows with the system.
02
< 1 day
to the root cause of a cross-domain failure — instead of weeks of expert coordination
Today, system knowledge is fragmented across domains, suppliers, and experts. PULSAR preserves a behavioral reference for every software version, so the understanding stays with the product as teams, suppliers, and experts change. The answer no longer depends on who is available.
03
No specifications to complete. No behavior to describe. Nothing new to instrument. PULSAR learns from the logs and traces your system already produces — and builds its first behavioral reference from data recorded before you ever spoke to us.
Get started
Give every release a memory. Give every failure a path.
