Due　to　the　problems　of　poor　scalability,　difficult　experimental　evaluation,　and　the　lack　of　teaching　data　analysis　and　collaborative　sharing　in　traditional　experimental　teaching　platforms,　this　paper　designs　an　interactive　and　scalable　intelligent　experimental　teaching　auxiliary　platform　BERTDS　based　on　deep　learning　algorithms　and　computer　technology.　The　platform　provides　a　wide　range　of　functions,　such　as　the　release　of　experimental　resources,　online　Q&A,　cloud　storage　sharing,　automatic　evaluation,　similarity　detection,　evaluation　and　assignment　management,　etc.　This　paper　first　introduces　the　design　idea　and　overall　architecture　of　the　experimental　platform　based　on　the　deep　learning　BERT　framework;　then　expounds　the　design　of　the　organization　module　and　automated　evaluation　engine　that　support　a　variety　of　experimental　schemes　and　the　distributed　deployment　scheme　of　the　server;　finally,　through　the　actual　application　data　analysis　and　user　Research　feedback　to　prove　the　feasibility　and　effectiveness　of　the　platform.　©　2023,　The　Author(s),　under　exclusive　license　to　Springer　Nature　Singapore　Pte　Ltd.