Consulting and workshops
For Organizations and Teams
Technical workshops, applied consulting, and literature sessions for research and engineering teams.
Workshops
Build a working understanding of inverse methods
I design technical lectures and workshops for engineering and research teams that need to evaluate, use, or direct the development of inverse-problem methods.
Electrical impedance tomography
Sensing hardware, the complete electrode model, reconstruction methods, and the limitations imposed by noise and partial boundary measurements.
Full waveform inversion and seismic imaging
PDE-governed forward models, adjoint-based gradients, regularization for velocity-model estimation, and common failure modes.
Ill-posed inverse problems
Why small data errors can produce large reconstruction errors, what regularization can achieve, and how to identify an unstable inversion workflow.
Formats range from a focused seminar to a one-day intensive or a short series. The intended outcome is practical technical judgment, not familiarity with terminology alone.
Consulting
Review the mathematical structure before committing to a pipeline
My consulting work begins with the forward model, measurement process, and noise structure. Only then do we assess an inversion method, computational strategy, or learned component.
Parameter identification
Recovering conductivity, velocity fields, source terms, or other model parameters from indirect measurements.
Pipeline review
Discretization, adjoint consistency, regularization, stopping criteria, and parameter-selection strategy.
Uncertainty analysis
Bayesian formulations, sensitivity analysis, credible intervals, and the limits of Gaussian approximations.
Scientific machine learning
Assessing whether PINNs, neural operators, or learned regularizers are appropriate for an existing workflow.
Working boundary
If a problem depends on a forward model outside my background, I will identify that limitation directly. The initial conversation is used to establish technical fit before any engagement.
Literature
Track research without mistaking novelty for readiness
I run recurring sessions for teams following scientific machine learning, inverse problems, and computational methods. Sessions examine recent papers, distinguish substantial advances from incremental results, and connect findings to the team's actual technical problems.
A monthly cadence is typical. This format works best for teams with existing computational infrastructure that need a rigorous view of which methods are mature enough to test or adopt.
Typical coverage
- PINNs for forward and inverse problems
- Neural operators and surrogate models
- Physics-constrained learning
- Learned regularization and deep unrolling
- Reproducibility and benchmark quality
- Evidence required for applied deployment
Contact
Start with the technical problem
Send a short description of the model, measurements, current workflow, and decision your team needs to make. The introductory call is free and takes 20–30 minutes.