Typical areas of engagement include:
Operations and Monitoring
Determining the health of a complex and inteconnected system requires
more than time-series analysis of logfile data. Instead, we need
mathematical models for the expected behavior in order to define
criteria to identify outliers and anomalies. Methods from statistical
process control then detect anomalous conditions early and reliably,
and trigger corrective actions.
Flow and Load Control
Adapting to variable loads and working conditions through auto-scaling
and active queue managment is a requirement for reliable operations.
Control theory and feedback designs have long been used in the engineering
professions to build reliable systems, but are almost unknown to software
I have written “the book” on control theory for computer systems.
If you have a control problem, I can help.
Optimization and Cost Control
Optimization problems, in particular combinatorial optimizations
(traveling salesman-type problems, such as routing, scheduling,
bin fitting problems) in general cannot be solved using textbook methods,
but require problem-specific approximation methods and heuristics.
(I once solved a client’s scheduling problem using methods from
Quantum Field Theory.)
Numerical analysis and computationally intensive programming projects?
Discrete event simulations? Defect reduction and preventive maintance?
Data model, dashboard, control tower design? Stuff too weird to describe?
Maybe we should talk…