I will audit your rag system and reduce hallucinations

R
royxforge
R
royxforge
Sourav Roy
Einige Informationen werden in englischer Sprache angezeigt.
audit your rag system and reduce hallucinations
Vollbild
audit your rag system and reduce hallucinations
Vollbild
Vollbild
audit your rag system and reduce hallucinationsaudit your rag system and reduce hallucinationsaudit your rag system and reduce hallucinations

Über diesen Service

I will audit your RAG system and show you exactly where it fails with numbers.


I'm an AI engineer with 3 research papers and production RAG systems in healthcare, education, and sales. I built an open-source evaluation framework that scores 6 dimensions: faithfulness, hallucination rate, retrieval precision, answer relevance, context coverage, and a label-free confidence metric (UCM) that needs zero ground-truth labels.


WHAT YOU GET:

- Evaluation run on your real queries (OpenAI, Anthropic, Llama, or any provider)

- Hallucination rate + faithfulness scores, ranked by severity

- Root causes: chunking, retrieval, prompts, reranking

- Prioritized fix roadmap


Standard adds a full 6-dimension benchmark report. Premium implements the fixes and re-runs evaluation to show before/after numbers.


Works with LangChain, LlamaIndex, FastAPI, vector DBs (Pinecone, Chroma, FAISS, Weaviate).


Send your repo/docs and I'll start within 24h.


  • Modell-Expertise

    • Entwicklung von individuellen Modellen
    • Feinabstimmung von Modellen
    • Generative KI
    • Prädiktive Analytik
    • Empfehlungssysteme

  • Industrie

    • Datenanalyse
    • Finanzdienstleistungen
    • Essen & Getränke
    • Recht
    • Sport & Fitness
    • Transport & Kfz
    • Reisen & Tourismus

  • Programmiersprache

    • C++
    • Java
    • Python
    • React
    • PyTorch
    • keras

  • Sprache

    • Bengalisch
    • Englisch
    • Hindi

  • Technische Expertise

    • Maschinelles Lernen (überwacht, unüberwacht, bestärkend)
    • Deep Learning (Neuronale Netze, GANs)
    • Computerlinguistik (CL)
    • Computer Vision (Objekterkennung, Bilderkennung)
    • Reinforcement Learning (Entscheidungssysteme)
    • Algorithmenentwicklung und -optimierung
    • Feature Engineering und Datenaufbereitung
    • KI-Ethik und Minderung von Bias

Lerne Sourav Roy kennen

Sourav Roy

AIML Engineer

  • AusIndien
  • Mitglied seitJuni 2026
  • ⌀ Antwortzeit1 Stunde

  • Sprachen

    Bengalisch, Hindi, Englisch, Italienisch
I am an AI/ML Engineer with three published research papers and experience building production systems in healthcare, education, and sales. I architect end-to-end pipelines involving LLM orchestration, RAG, and neural model calibration. My focus is on creating robust, scalable AI infrastructure that bridges the gap between applied research and real-world reliability.

Verwandte Tags