Browse categories
Explore
Fiverr Pro
English
$
USD
I will add ai observability, tracing and output evaluation to your app
Ian Ye
About this gig
Already using AI but missing visibility into calls and output quality?
I will add locally hosted Langfuse tracing and configurable evaluation to ONE existing Python or Node.js AI workflow.
Langfuse and its tracing data run on your own computer or private server, not Langfuse Cloud.
Includes:
- Review the workflow and modify integration code
- Set up local Langfuse and trace agreed steps and model calls
- Record permitted inputs, outputs, errors, latency and available usage/cost estimates
- Configure agreed business checks and your evaluation model
- Export agreed input/output fields as structured JSONL for training-data preparation
- Deliver source changes, setup notes and a verified test workflow
Exports require your data-use permission and review before training. Model training, dataset curation, custom dashboards, extra workflows and new evaluation algorithms are separate.
You provide source access, a working test environment, samples and criteria, and pay hosting/model fees. Evaluation uses your approved local or external endpoint; external evaluation sends agreed data to that provider.
Message me before ordering to confirm scope and acceptance criteria.
Get to know Ian Ye
Ian Ye
AI Product Builder, Product Design, Web Apps and Workflows
- FromChina
- Member sinceJul 2026
- Avg. response time17 hours
Languages
English
I build web apps and AI tools, combining 5 years in product management with 3 years in frontend development.
My projects span e-commerce, content creation, learning, and model API platforms. I offer requirements analysis, user flow design, AI-assisted development, API integration, functional testing, and AI output evaluation.
As a native Chinese speaker, I also help review Chinese AI responses, improve Simplified Chinese UI copy, and check localized user flows.
I work through written communication, with clear scope and deliverables agreed upfront.
My Portfolio
Other AI Development Services I Offer
FAQ
What do you need before starting?
I need access to the relevant source code, setup instructions, a working test environment, the target AI workflow, representative samples and your evaluation criteria. Please provide credentials through an agreed secure method.
Does my app need LangChain or an existing Langfuse integration?
No. I can work with existing Python or Node.js AI applications, including direct model API calls. LangChain is not required. I will review your source code first to confirm compatibility and the required changes.
Can evaluations follow my own business rules?
Yes. We will define checks, scoring criteria, thresholds and reference material for your workflow. Rule configuration will be documented. New custom algorithms or a visual rule editor require separate scope.
Will my tracing data be stored in Langfuse Cloud?
No. Langfuse will run in your own environment. Model-based evaluation uses the endpoint you approve, which may be local or external. Self-hosted tracing does not automatically mean local model inference.
What will I see after an AI call?
You can inspect captured inputs, outputs, execution steps, errors, latency and available usage data in Langfuse. Agreed evaluations show scores or check results with reasons. Missing usage or cost data will be marked as unavailable.
Does this package include a custom dashboard or training dataset?
This package uses the self-hosted Langfuse interface. A separate dashboard, human review system, curated training dataset or model training is not included. These can be scoped separately after reviewing your requirements.
Are hosting, model fees and additional features included?
Hosting and model API fees are paid by you. One revision covers adjustments within the agreed workflow and acceptance criteria. Additional workflows, new business features or changes to the agreed scope require a separate offer.
