Kategorien durchsuchen
Entdecken
Fiverr Pro
Deutsch
$
USD
I will build a custom llm evaluation and testing framework
Zlatin S
Vetted Pro
Level 1
Geprüft von Fiverr Pro
Zlatin S wurde vom Fiverr Pro-Team aufgrund besonderer Expertise ausgewählt.
Über diesen Service
Your AI app works today. Will it still work after a model or prompt update?
I build Python-based LLM evaluation tools that help teams detect failures, track response quality, and catch regressions before release.
WHAT I BUILD
- Automated tests for LLM APIs and chatbots
- Output-format and expected-behavior checks
- Rubric-based response scoring (Standard and Premium)
- Reports showing failed cases and key metrics
- Streamlit results dashboard
- Model or prompt comparison and GitHub Actions CI (Premium)
PACKAGES
Basic: 25 test cases, 2 checks, and a simple results dashboard.
Standard: 75 test cases, 3 criteria, detailed analysis dashboard.
Premium: 150 test cases, 4 criteria, comparison and CI automation.
YOU RECEIVE
Python source code, evaluation test files, result exports, setup instructions and a demonstration.
Scope: One agreed English-language workflow. Model API charges, hosting, fine-tuning and ongoing monitoring are excluded. No evaluation system guarantees zero hallucinations.
Please message me before ordering to confirm your API, test data and acceptance criteria.
Lerne Zlatin S kennen
Zlatin S
Vetted Pro
Senior AI and Full Stack Engineer
Level 1
Zlatin S ist Teil des Fiverr Pro-Katalogs und wurde von einem spezialisierten Fiverr Pro-Team basierend auf Fähigkeiten und Fachwissen sorgfältig ausgewählt.
Geprüft für
KI-Entwicklung
- AusBulgarien
- Mitglied seitAug. 2025
- ⌀ Antwortzeit4 Stunden
- Letzte Lieferung2 Monate
Sprachen
Bulgarisch, Englisch
I'm a Senior Full-stack AI engineer with over 10 years of experience so I focus on developing AI-powered SaaS platforms, Web, and Mobile applications.
My focus is turning manual processes into scalable systems that save time and increase revenue.
I specialize in
- AI agents & LLM integrations (RAG, Chatbot, OpenAI, Claude, custom pipelines),
- AI-powered SaaS platforms (from backend to UI),
- workflow automation (CRM, support, internal tools),
- data pipelines and intelligent systems,
- Expert in LLM integration, AI agents, Full-stack development, and end-to-end product delivery.
Mein Portfolio
Meine weiteren Dienstleistungen im Bereich KI-Entwicklung
FAQ
What types of AI applications can you evaluate?
I can build tests for text-based LLM APIs, chatbots, and similar AI applications. More complex agent or multimodal workflows require a custom scope.
Which LLM providers do you support?
I support OpenAI-compatible endpoints and other providers with accessible APIs, subject to confirming their integration requirements before the project.
Can you detect AI hallucinations?
I can implement groundedness and reference-based checks where suitable data is available. These checks help identify potential hallucinations but cannot guarantee perfect detection.
Can you test a RAG chatbot?
Yes, within an agreed response-evaluation scope. Advanced document retrieval testing, citation verification, and end-to-end RAG testing may require a custom offer.
Will I receive the complete source code?
Yes. I provide the agreed Python source code, test cases, evaluation reports, and setup instructions.
Can I run the tests after delivery?
Yes. The framework is designed for repeated execution using your configured API access and local environment.
Are API usage and hosting included?
No. Third-party model usage, API credits, cloud hosting, and ongoing monitoring are not included in the package price.
What is included in a revision?
A revision covers corrections and adjustments within the agreed specifications. New models, integrations, or major features require a separate quote.
