Gene finding plays a fundamental role in bioinformatics, the fastest growing area of computational science. Accurate gene calling is crucial for a success of a number of computational and experimental approaches developed to elucidate molecular mechanisms of life. Comparative genomic characterization of protein function, manufacturing DNA microarrays, building biomolecular networks rely on data on the whole complement of genes identified in a given genome by gene finding tools. Although Georgia Tech does not have a genome center, the gene finding algorithms developed at Georgia Tech were of primary use in five out of the top ten genome projects. I will review the major algorithms that we have developed. These algorithms use Markov and Hidden Markov models with parameters estimated from genomic data. In our latest developments for both prokaryotic and eukaryotic genomes we have implemented unsupervised machine leaning for estimation of model parameters. Our latest algorithms allow accelerating a tedious process of genome annotation, particularly fungal and plant genomes, by at least several months and are now in use in many research labs and genome sequencing centers including DOE Joint Genome Institute, J. Craig Venter Institute and Broad Institute.