The main way this is used in portfolio construction is that you start with an index fund portfolio, i.e. with weights that are proportional to market caps. An index fund will usually have a market beta of 1 and SMB and HML betas that are fairly small (close to zero). You then try to enhance the SMB and HML betas of the portfolio, by increasing (though some mathematical formula) the weights of stocks with high SMB and HML beta and reduce the weights of stocks with low SMB and HML beta.
In this way you arrive at a portfolio that is still well diversified, has no short positions and tracks the index fairly closely (with market beta of 1 and modest Tracking Error) but has the desired "overloading" of HML and SML. People who do this believe that this extra exposure will be beneficial for long run performance.
Probably the earliest firm to do this was Dimensional Fund Advisors, but nowadays many more firms do it (and not only with SMB and HML as factors, but others as well) and it has generally become known as Factor Investing. As always when many people do something in investing, there are questions as to whether it will continue to be a good idea.