for someone who wants to try out building a conversational ai setup with nvidia products, is nvidia NeMo still the default place to start like is it still relevant? or theres some other start point that I should look into to build out a conversational ai setup, looking for suggestions here