Expertise

Our staff has in-depth expertise in a diverse set of industries and an intimate knowledge of tools required for software, data, and algorithm research and development. These include:

  • MATLAB toolboxes: Econometrics, Neural networks, Image processing, Parallel processing, Curve fitting
  • C/C++/C# and Java
  • LabView
  • Database tools: tSQL, mySQL, Python, Perl, Shell Scripts
  • Mathematica

Some of our recent projects included:

  • Hyperspectral imaging
  • Continuous glucose monitoring and insulin delivery
  • Mobile and web analytics

Some of our clients work in industries that are subjected to regulatory oversight. While we do not have direct expertise in regulatory compliance, we a sensitive to these clients' needs for enchanced documentation and quality processes specific to their work.

Data Collection and Visualization

Data comes in many shapes and sizes. Whether your data is stored in a database, flat file, or a proprietary format, we can retrieve, process, and visualize it in a meaningful way. We are experienced with:

  • tSQL/mySQL for accessing Database on a server
  • Direct database access from MATLAB
  • LabView for real-time data collection from a sensor network
  • Various imaging formats including hyperspectral imaging. Image acquisition directly from a camera system
  • Tableau reporting
  • API calls over HTTP

and other data formats.

Feedback, Filtering and Optimization

We are familiar with the tools, mathematics, and application of standard and advanced optimization and filtering techniques:

  • Filter design: FIR, PID, State Space feedback controller design
  • LMS, Kalman, and other filters
  • LabView and/or Simulink in-process integration
  • Model based predictive control
  • Constrained optimization
  • Portfolio optimization

Modeling and System Identification

Once the data is retrieved and pre-processed, we can use it to build a mathematical model of your system's behavior.  We will estimate the model's parameters and test its accuracy based on historical or simulated data. Then, we will use this representation of your system to infer its future behavior (forecasting) and to separate meaningful signals from noise (filtering).  Our core modeling strengths include:

  • Frequency response-based system identification
  • Wavelet analysis
  • Principal component analysis and linear discriminate analysis
  • Non-parametric methods including Kernel Density Estimators
  • Look up table : Numerical, polynomial or NURBS interpolation
  • Neural network learning; genetic algorithms
  • Bayesian statistical learning and pattern classification
  • Seasonally adjusted ARIMA estimation and forecasting
  • Monte Carlo modeling of time-series (particularly for financial or web analytics applications)
  • Solution to integro-differential equations using Monte Carlo methods, especially in the context of diffuse light transfer
  • ODE and PDE modeling of dynamic systems

Making a Product

After developing and successfully testing the algorithm, we integrate it into your product's lifecycle. This may include productizing the algorithm a language such as C# or Java. This work can be done in-house or in collaboration with your software engineering team. Finally, we provide technical training and complete documentation to your staff to hand over your product.