Rank Team run_ID Baseline ID errors MAP_binary MAP_graded MRR_binary MRR_graded NDCG_binary NDCG_graded P@5_binary P@5_graded P@10_binary P@10_graded P@100_binary P@100_graded Representation type Methodology System details LLM usage Bias mitigation External data Models Pipeline Representation LLM strategy Fine-tuned Uses external data Bias handling 1 classum 715319 no no 0.4197 0.4197 0.9077 0.9077 0.834 0.8068 0.7692 0.7692 0.7368 0.7368 0.5149 0.5149 Encoder; Decoder Retrieval; Reranking Batch contrastive loss; CachedGISTEmbedLoss; GISTEmbedLoss; Encoder Fusion Prompt engineering; Data Augmentation Yes ZEmbed1-4B; KaLM-12B; Qwen3-8B; NVEmbed-3B; Zerank2 Retrieval+Reranking Dense+Generative Augmentation Yes Yes No 2 classum 716006 no no 0.4197 0.4197 0.8977 0.8977 0.8338 0.8067 0.7795 0.7795 0.735 0.735 0.5149 0.5149 Encoder; Decoder Retrieval; Reranking Batch contrastive loss; CachedGISTEmbedLoss; GISTEmbedLoss; Encoder Fusion Prompt engineering; Data Augmentation Yes ZEmbed1-4B; KaLM-12B; Qwen3-8B; NVEmbed-3B; Zerank2 Retrieval+Reranking Dense+Generative Augmentation Yes Yes No 3 classum 710704 no no 0.4087 0.4087 0.8987 0.8987 0.8278 0.7986 0.7453 0.7453 0.7026 0.7026 0.5103 0.5103 Encoder; Decoder Retrieval; Reranking Batch contrastive loss; CachedGISTEmbedLoss; GISTEmbedLoss; Encoder Fusion Prompt engineering; Data Augmentation Yes ZEmbed1-4B; KaLM-12B; Qwen3-8B; NVEmbed-3B; Zerank2 Retrieval+Reranking Dense+Generative Augmentation Yes Yes No 4 NightSun 709648 no no 0.4089 0.4088 0.8213 0.8213 0.8246 0.7913 0.6991 0.6991 0.6761 0.6761 0.5081 0.5081 Encoder; Decoder Retrieval; Reranking GISTEmbedLoss Pointwise reranking; Pairwise reranking; Tournament pairwise reranker; Hypothetical Document Embeddings Query Expansion No JobBERT-v2; Qwen-2.5-7B-Instruct-AWQ Retrieval+Reranking Dense+Generative Reranking + augmentation Yes No No 5 NightSun 708235 no no 0.4005 0.4004 0.8137 0.8137 0.8209 0.7866 0.7009 0.7009 0.6726 0.6726 0.4626 0.4626 Encoder; Decoder Retrieval; Reranking GISTEmbedLoss Pointwise reranking; Pairwise reranking; Tournament pairwise reranker; Hypothetical Document Embeddings Query Expansion No JobBERT-v2; Qwen-2.5-7B-Instruct-AWQ Retrieval+Reranking Dense+Generative Reranking + augmentation Yes No No 6 bipboopbipboop 714675 no no 0.3889 0.3889 0.8882 0.8882 0.8123 0.7793 0.7385 0.7385 0.7034 0.7034 0.5137 0.5137 Encoder; Decoder Retrieval; Reranking Query expansion; InfoNCE; LoRA; Score Fusion Listwise select-then-rank; Pointwise reranking; Prompt engineering No Qwen3-Embedding-4B; Qwen3-Reranker-8B; Qwen3-235B-A22B-Instruct; JobBERT-v3 Retrieval+Reranking Dense+Generative Reranking + augmentation Yes No No 7 bipboopbipboop 714691 no no 0.3799 0.3799 0.8996 0.8996 0.8095 0.7766 0.7487 0.7487 0.7256 0.7256 0.4963 0.4963 Encoder; Decoder Retrieval; Reranking Query expansion; InfoNCE; LoRA; Score Fusion Listwise select-then-rank; Pointwise reranking; Prompt engineering No Qwen3-Embedding-4B; Qwen3-Reranker-8B; Qwen3-235B-A22B-Instruct; JobBERT-v3 Retrieval+Reranking Dense+Generative Reranking + augmentation Yes No No 8 NightSun 708252 no no 0.3572 0.357 0.7617 0.7617 0.7957 0.7482 0.5829 0.5829 0.5376 0.5376 0.4325 0.4325 Encoder; Decoder Retrieval; Reranking GISTEmbedLoss Pointwise reranking; Pairwise reranking; Tournament pairwise reranker; Hypothetical Document Embeddings Query Expansion No JobBERT-v2; Qwen-2.5-7B-Instruct-AWQ Retrieval+Reranking Dense+Generative Reranking + augmentation Yes No No 9 hr_gradient 716124 no no 0.3093 0.3093 0.8052 0.8052 0.7746 0.7399 0.653 0.653 0.6043 0.6043 0.4134 0.4134 Encoder Retrieval; Reranking Cross-encoder reranking; Domain adaptation; CachedGIST Prompt Engineering; Data Augmentation Yes Alibaba-NLP/gte-modernbert-base; Qwen/Qwen3-Reranker-0.6B; Gemma 4 Retrieval+Reranking Dense Augmentation Yes Yes No 10 bipboopbipboop 705971 no no 0.2962 0.2962 0.883 0.883 0.751 0.7294 0.7675 0.7675 0.7094 0.7094 0.4167 0.4167 Encoder; Decoder Retrieval; Reranking Query expansion; InfoNCE; LoRA; Score Fusion Listwise select-then-rank; Pointwise reranking; Prompt engineering No Qwen3-Embedding-4B; Qwen3-Reranker-8B; Qwen3-235B-A22B-Instruct; JobBERT-v3 Retrieval+Reranking Dense+Generative Reranking + augmentation Yes No No 11 Skillberg 710631 no no 0.2736 0.2736 0.6451 0.6451 0.7487 0.7088 0.494 0.494 0.4769 0.4778 0.3944 0.3944 Encoder Retrieval; Classifier Proprietary job-skill knowledge graph; Cached InfoNCE No intfloat/multilingual-e5-large-instruct Retrieval Dense None Yes No No 12 Skillberg 708268 no no 0.2634 0.2634 0.6228 0.6228 0.7427 0.7001 0.4701 0.4701 0.4641 0.465 0.3623 0.3623 Encoder Retrieval; Classifier Proprietary job-skill knowledge graph; Cached InfoNCE No intfloat/multilingual-e5-large-instruct Retrieval Dense None Yes No No 13 baorphuc 692732 no no 0.2359 0.2359 0.7017 0.7017 0.7285 0.6829 0.4872 0.4872 0.4701 0.4701 0.325 0.325 Encoder Retrieval InfoNCE No all-mpnet-base-v2 Retrieval Dense None Yes No No 14 ui-nlp 716152 no no 0.2366 0.2366 0.6297 0.6297 0.7278 0.6821 0.453 0.453 0.4299 0.4291 0.3291 0.3291 Encoder Retrieval InfoNCE No intfloat/multilingual-e5-large-instruct Retrieval Dense None Yes No No 15 MARSAD 697840 no no 0.1812 0.1812 0.6508 0.6508 0.695 0.6531 0.5094 0.5094 0.4385 0.4385 0.267 0.267 Encoder; Lexical Retrieval; Rank fusion No BM25; intfloat/e5-base-v2 Retrieval Dense+Lexical None No No No 16 Olive 693022 no no 0.1853 0.1853 0.5338 0.5339 0.6899 0.6458 0.3282 0.3282 0.3197 0.3188 0.2663 0.2664 Encoder; Lexical Retrieval Curriculum Learning; InfoNCE; Reciprocal Rank Fusion No BM25; BAAI/bge-base-en-v1.5; paraphrase-multilingual-mpnet-base-v2 Retrieval Dense+Lexical None Yes No No 17 TalentCLEF baseline 690199 yes no 0.1471 0.1471 0.5701 0.5701 0.6634 0.6204 0.3179 0.3179 0.3043 0.3017 0.2308 0.2308 18 MARSAD 691629 no no 0.1471 0.1471 0.5701 0.5701 0.6634 0.6204 0.3179 0.3179 0.3043 0.3017 0.2308 0.2308 Encoder; Lexical Retrieval; Rank fusion No BM25; intfloat/e5-base-v2 Retrieval Dense+Lexical None No No No 19 Skillberg 690315 no no 0.2612 0.2612 0.6947 0.6947 0.6384 0.6129 0.5179 0.5179 0.5017 0.5017 0.3847 0.3847 Encoder Retrieval; Classifier Proprietary job-skill knowledge graph; Cached InfoNCE No intfloat/multilingual-e5-large-instruct Retrieval Dense None Yes No No