Browse categories
Explore
Fiverr Pro
English
$
USD
I will build a custom rag ai chatbot for your website
krishna k
_gp0nvd.jpg)
About this gig
I build custom AI chatbots that answer questions grounded in your actual data documents, product catalogs, policies, FAQs using Retrieval-Augmented Generation (RAG), not a generic ChatGPT wrapper that makes things up.
What I've built:
An internal knowledge assistant for a company's employee and policy documents, with source citations and conversational follow-up handling
A customer-facing shopping assistant for an e-commerce brand that accurately answers product, sizing, and live inventory questions
My approach:
Every chatbot uses a real retrieval pipeline chunking, embeddings, vector search (ChromaDB), query expansion, and cross-encoder reranking so answers are accurate and traceable, not hallucinated.
I also test retrieval quality with structured evaluation (MRR, nDCG, keyword coverage, and accuracy scoring across 50-100 test questions) so you get proof your chatbot works, not just a demo.
Tech stack: Python, LangChain, ChromaDB, HuggingFace Embeddings, OpenAI GPT, FastAPI, Cross-Encoder Reranking.
Use cases: Customer support, e-commerce product/order assistance, internal knowledge bases.
Message me before ordering if your use case doesn't fit a package exactly happy to scope it
Get to know krishna k
krishna k
AI Chatbot Developer specializing in RAG Systems, LangChain and OpenAI
- FromIndia
- Member sinceAug 2026
- Avg. response time1 hour
Languages
Hindi, English
I build AI chatbots that answer from your actual
business documents — PDFs, manuals, FAQs, websites.
Using RAG, your chatbot reads real documents and cites
sources with every answer — not making things up.
What I build:
- Document-trained chatbots
- Conversation memory & follow-ups
- Embeddable website chat widgets
- Tool integrations & AI agents
I measure accuracy using MRR
and NDCG before delivery — proven to work, not just
built and hoped for the best.
Message me and I'll tell you honestly if I can help.
My Portfolio
FAQ
Do I need any technical knowledge to use this?
No. I'll set everything up and hand over a working chatbot with simple instructions — no coding required on your end.
Can it work with my own documents/data?
Yes. I can build the chatbot using your PDFs, product catalogs, policies, FAQs, or other documents as the knowledge source.
What is RAG and why does it matter?
RAG (Retrieval-Augmented Generation) means the chatbot looks up real information from your documents before answering, instead of guessing. This makes answers accurate and reduces hallucinations — a common problem with basic ChatGPT-wrapper bots.
How do I know the chatbot is actually accurate?
I offer an optional evaluation report (Gig Extra) where I test the chatbot on 50-100 real questions and score it on retrieval and answer-quality metrics — so you get measurable proof, not just a demo.
Can I see the source code?
Yes, source code is included in Standard and Premium packages (available as an add-on for Basic).
What if my use case doesn't fit a package exactly?
Message me before ordering — I'm happy to scope out a custom solution if your project needs something different.
Do you offer revisions?
Yes, each package includes revisions (see package details), with additional revisions available as an extra.
What platforms does the chatbot work on?
Currently I build website-embedded chatbots. Other platform integrations may be available on request — message me to discuss.
