I will compare china llms for your task and deliver a bakeoff PDF

N
nan_lin_dev
N
nan_lin_dev
NanLin

About this gig

Need to know if DeepSeek, Qwen, Kimi, or GLM can replace GPT/Claude for YOUR workflowwithout guessing from leaderboards?


I run a China LLM cost × quality bake-off on your task (invoice extract, lead scoring, or a similar structured job) and deliver:

Decision PDF most accurate / cheapest (above pass threshold) / best value

Evidence appendix failed or disputed sample × model rows


How it works

1) Kickoff: lock prompt + scoring rules (you approve in writing)

2) Samples: your labeled set (best) or a small agreed set

3) Run the same cases on 4 domestic models; optional GPT/Claude baseline via your key or frozen outputs

4) You get A + B PDFs ready to share with stakeholders


This is NOT chatbot development, RAG build, fine-tuning, or source-code delivery. Category limits applythe product is a bake-off report.


You provide: task definition, samples (no real PII in first pass if possible), and whether baseline is client key or frozen outputs.

I provide: scoring-spec draft, eval run, decision PDF, evidence appendix.


Start with Basic for a single-task sample bake-off; upgrade when you have gold labels.

Get to know NanLin

NanLin

n8n Automation AI Integration Specialist

  • FromChina
  • Member sinceMay 2026
  • Languages

    English, Chinese
Hi, I’m Nan. I build production-ready automation with n8n, Python, Docker, webhooks, and AI — focused on B2B lead pipelines: forms → AI scoring → Google Sheets → Slack → HubSpot DRAFT → Calendly + digests, with Langfuse/Jaeger. I don’t just connect APIs. I design for error handling, retries, test→production gates, and clear docs so systems stay maintainable and cost-efficient. Also experienced with Make, Apps Script, and RSS/Discord/Telegram alerts. Need lead automation, CRM sync, or a reliable n8n system? Message me your stack — I’ll recommend the right scope.

My Portfolio