The best machine learning consultants?

I can comment on the two ways I've been found, which is mostly applicable for finding grad student consultants.

(1) People have searched the proceedings of conferences and emailed me off of those papers (2) LinkedIn. Join either the machine learning or NLP groups. (3) Graduate student directories. Me and several people from my old department used to get the same recruitment email. It was clear whoever sent it, sent the same request to everyone who listed "Machine learning" as an area of concentration

Best, Aria

There are two issues here: the speciality of the consultant, and that the consultant is good.

You want a consultant who focuses on the sort of issues that interest you. If integration with existing enterprise technology is a key concern, you should consider someone with a lot of industry experience. Or, you might want something state-of-the-art, in which case you should get someone fresh out of academia who is up with the latest techniques.

To get someone who is very good at this focus, I think referrals and recommendations are the best route. This is true for almost all hiring.

I run the A.I.Cookbook which puts out recipes for easy A.I. The Google Group has a lot of interesting people from the industry, there's a reasonable chance you'll find someone who can help in this Group.

I'm an A.I. consultant too specialising in evolutionary optimisation, natural language processing and ways of pulling hard problems apart to figure out which problems really need to be tackled. More at Mor Consulting.

Doesn't hurt to do a web or Linked In search for "consulting" or "consultant" and the specialty you're interested in.

But I agree that referrals are the best way to find someone good. Concretely, that means identifying the broad area of the problem, and contacting people who have worked in that area to ask if they know a consultant they think highly of, or know someone who might know someone.

It's important to have a clear sense of the kind of person you're looking for (e.g. grad student who's a good developer, distinguished expert to testify in a patent case, etc.), but less important to know how the problem is to be solved. The latter is what you're hiring the consultant for.

University faculty are good sources, since their publications give you a sense of their area(s). Some do consulting on the side, and many who don't will know people who do.

Regards, Dave

p.s. My own technical areas are information retrieval, machine learning, natural language processing, and effectiveness evaluation for complex information processing systems. Industries I've consulted to include web search, e-discovery, computational advertising, information security, data mining, venture capital, biotech, defense, nonprofits, and government. More at www.DavidDLewis.com.