About Sondera
Inspection should end in a decision you can defend.
Every year, industry captures more inspection data and understands less of it. Sondera exists to close that gap: to make the data behind every safety-critical part fully understood, not just recorded, across every nondestructive evaluation modality.
Our mission
From signal to decision
To turn raw inspection data into trustworthy, traceable decisions, with AI that stays grounded in the physics of the measurement, so NDE professionals can act on evidence at the scale their work demands.
What we believe
Four commitments we do not trade away.
01
Physics before pattern-matching
Models are grounded in the wave physics of each modality, not fit blindly to data. Where a measurement standard applies, there are no black boxes.
02
Traceable by default
Every decision links back to the raw data, the model version, and the acceptance criterion applied. Engineers sign off on evidence, not intuition.
03
Honest about uncertainty
Every call carries a confidence measure. A borderline indication stays borderline instead of being quietly rounded to pass or fail.
04
Your data stays yours
We process the data you already have and return results and models you can audit. No lock-in, no quiet reuse of your inspections.
How we work
Physics first, then machine learning, then a report.
Simulate first
Physics-based simulation, FDTD and finite element, models the measurement before hardware time is spent and generates training data grounded in real wave behavior.
Learn with physics in the loop
Physics-informed models are trained and benchmarked against simulated and measured reference data, so performance is known rather than assumed.
One pipeline, every modality
The same signal-to-decision pipeline serves terahertz today and extends to ultrasonic and micro-CT, so each new modality reuses a proven core.
Fit your workflow
Sondera works with the instruments and data you already have, and returns auditable reports that match how your team already signs off.
Built on research
The roadmap is a research agenda, not a wish list.
Sondera grows out of doctoral and peer-reviewed research in terahertz nondestructive evaluation and physics-based modeling of composites. That work already spans the three modalities on our roadmap: terahertz imaging, X-ray micro-computed tomography, and multi-sensor fusion. Every method is validated against simulation and reference measurements before it touches a real part.
- Deep-learning terahertz damage characterization in composite laminates
- Published in peer-reviewed venues including Composites Part B: Engineering
- Micro-CT damage detection research under review at NDT & E International
- Physics-informed models benchmarked against reference measurements
Leadership
The people behind Sondera.
Niyem Mawenbe Bawana
Founder
Sondera was founded by Niyem Mawenbe Bawana, whose doctoral research is the technical foundation of the platform. His PhD at the Georgia Institute of Technology develops deep learning for terahertz damage characterization in composite laminates, the exact problem Sondera solves first. His research reaches from applied physics through electrical and mechanical engineering, the multiphysics grounding that keeps Sondera's models tied to measurement reality. A Fulbright scholar, he joins the University of Colorado Boulder as an Assistant Teaching Professor in 2026, and has mentored engineering students on nondestructive evaluation and sensing projects for years.
- PhD, Georgia Institute of Technology
- Fulbright Scholar
- Peer-reviewed terahertz NDE research
- Incoming faculty, CU Boulder
A growing team
Sondera is early and building deliberately. Alongside a network of research collaborators, we are bringing on interns and volunteer contributors in signal processing, physics simulation, and platform engineering, and we are honest that these are unpaid roles for now. The people who help build the foundation are the first we want to grow with.
See open rolesWork with us, or help build what comes next.
Bring us an inspection problem, join the THz pilot, or talk to us about sponsoring the work.