This paper introduces RAGnaroX, a resource-efficient ChatOps assistant that operates entirely on commodity hardware. Unlike existing solutions that often rely on external providers such as Azure or OpenAI, RAGnaroX offers a fully auditable, on-premise stack implemented in Rust. Its architecture integrates modular data ingestion, hybrid retrieval, and function calling, enabling flexible yet secure deployment. Our evaluation focuses on the RAG pipeline, with benchmarks conducted on the SQuAD (single-hop QA), MultiHopRAG (multi-hop QA), and MLQA (cross-lingual QA) datasets. Results show that RAGnaroX achieves competitive accuracy while maintaining strong resource efficiency, for example, reaching 0.90 context precision on single-hop questions with an average response time of 2.5 seconds per request. A replication package containing the tool, the demonstration video (https://www.youtube.com/watch? v=cDxfuEbcoM4), and all supporting materials are available at https://github.com/genius-itea/RAGnaroX.git.
Figures & tables
Fig. 1: Conceptual overview of the RAGnaroX data integration and retrieval–generation pipeline for ChatOps.
Fig. 2: Performance Metrics in RAGnaroX, with changed generation models
Model
Single-hop QA
Multi-hop QA
Faith.
AnsRel.
Faith.
AnsRel.
Qwen3 (14B)
0.8327
0.7846
0.7039
0.6407
Qwen3 (8B)
0.8364
0.7865
0.6874
0.6328
Mistral (7B)
0.7963
0.7705
0.4654
0.6568
Gemma-3n-E2B
0.7183
0.7722
0.4067
0.5056
Qwen3 (4B)
0.8588
0.8168
0.6341
0.6790
TABLE I: Faithfulness and answer relevancy comparison
Corpus– Question
Ctx. P.
Ctx. R.
Faith.
AnsRel.
14B
4B
14B
4B
en–en
0.86
0.91
0.83
0.81
0.73
0.75
de–de
0.74
0.77
0.77
0.73
0.41
0.44
es–es
0.82
0.87
0.77
0.80
0.53
0.58
en–de
0.59
0.70
0.74
0.70
0.41
0.43
de–en
0.64
0.73
0.73
0.71
0.71
0.65
TABLE II: RAG system performance of Qwen3 14B and 4B
This paper presents a highly efficient Retrieval-Augmented Generation (RAG) system built specifically for Ukrainian document question answering, which achieved 2nd place in the UNLP 2026 Shared Task. Our solution features a custom two-stage search pipeline that retrieves relevant document pages, paired with a specialized Ukrainian language model fine-tuned on synthetic data to generate accurate, grounded answers. Finally, we compress the model for lightweight deployment. Evaluated under strict computational limits, our architecture demonstrates that high-quality, verifiable AI question answering can be achieved locally on resource-constrained hardware without sacrificing accuracy.
Mykola Trokhymovych, Yana Oliinyk, Nazarii Nyzhnyk
Public institutions hold large volumes of sensitive documents and support tickets that cannot leave the premises, ruling out cloud-hosted language models entirely. We report on RAGAL, a retrieval-augmented assistant for the technical-support team of AFIR, the Romanian Agency for Financing Rural Investments, built and operated under three hard constraints: zero data egress (no external API calls, even for synthetic data), a read-only mandate (the assistant drafts, humans execute), and a single 8 GB consumer laptop as the only development and training machine. Over a Romanian-language corpus of ~25,000 chunks -- 15,073 resolved support tickets and internal normative documents -- we show that the highest-leverage investments were retrieval engineering and retriever fine-tuning rather than a larger generator: hybrid dense-sparse retrieval with intent routing raised our internal evaluation from 62% to 81%, and fine-tuning the bge-m3 embedder on real ticket data improved recall@10 from 0.663 to 0.850 (MRR 0.489 to 0.684) after 72 minutes of training. We document a general pitfall: single-domain fine-tuning silently degraded retrieval on the untouched document domain below the stock baseline, detected only after building a per-domain evaluation set and repaired with locally generated queries (GenQ). We report two counter-intuitive findings -- PII masking improved generation quality, and a structural "anchor distillation" scheme made SQL hallucination impossible by construction -- along with a reproducible recipe for full embedder fine-tuning in 8 GB of VRAM. Finally, since zero egress also rules out a cloud judge, we describe a substitute: a 744B-parameter model run on CPU, too slow to serve interactively but affordable in overnight batch, used as a second opinion whose limits we quantify. We release the sanitized pipeline scripts for institutions facing similar data-locality constraints.
Dan Musetoiu
IT Department, Agency for Financing Rural Investments (AFIR), Romania
While large language models have been dominating the research landscape recently, small language models remain highly relevant across various domains; yet, they receive far less attention. In this study, we investigate how smaller language models perform during the generation stage within a Retrieval-Augmented Generation (RAG) system. To benchmark these models effectively, we utilised both open-source and proprietary datasets covering diverse subject areas and question types. Our findings demonstrate that a RAG system with small language models can be executed directly on-device without requiring any GPU hardware within a reasonable time. The experimental code and links to the supplementary materials can be accessed through the GitHub repository: https://github.com/SibNN/SLM-RAG-EVAL.