Sfoglia categorie
Esplora
Fiverr Pro
Italiano
$
USD
I will build a custom rag chatbot using langchain
Sneh V
Informazioni su questo servizio
Most "AI agent" builds break the second something unexpected happens a bad input, a timeout, an edge case nobody tested for. I build agents that don't fall over.
I'm an AI engineer (IIT Kharagpur) who has built multi-step agent pipelines with retry logic, conditional routing, and error handling the kind of engineering that keeps things running when a single prompt-wrapper would just crash.
What I can build for you:
- A single-purpose agent that automates one specific task (data extraction, research, Q&A, scheduling, etc.)
- Multi-step agents that chain actions together search, verify, act
- Agents connected to your database or tools, not just a chat window
- Error handling so the agent recovers instead of silently failing
How it works:
- You tell me what task you want automated
- I build and test the agent against real inputs, including edge cases
- You get a working agent plus a short explanation of how it works
If you're tired of "AI agents" that only work in the demo, let's build one that works when it matters.
Message me with your use case and I'll tell you exactly what's possible.
Scopri di più su Sneh V
Sneh V
Creator
- DaIndia
- Membro damag 2025
Lingue
Inglese
I build AI systems that work in production, not just in a notebook. That means RAG chatbots that cite sources correctly, LLM fine-tuning pipelines that improve on baseline, and agent workflows that don't fall over the first time a rate limit hits.
I've trained transformer models from scratch and built retrieval systems that answer questions from thousands of documents.
Focus areas: RAG chatbots, LLM fine-tuning (DPO/LoRA), and AI agent pipelines.
GitHub: github.com/Snevj
Got a document, workflow, or model problem? Message me.
Il mio portfolio
FAQ
How is this different from just using ChatGPT with a prompt?
A prompt-wrapper answers one question and stops. An agent takes multi-step action — it can search, check its own output, retry if something fails, and only stop once the task is actually done.
What if my task changes or the agent hits something it wasn't built for?
That's exactly what the error handling covers. I test against edge cases before delivery, and revisions are included so we fix anything that comes up once you're using it for real.
Do I need to know how to code to use this?
No. You tell me the task, I build and test it, then hand it over with a plain explanation of how it works. No coding needed on your end.
What tools/platforms do you build agents on?
LangGraph, LangChain, and direct API integrations, depending on what fits your task. I'll tell you the right approach after hearing what you need.
Can the agent connect to my database or existing tools?
Yes, that's included in the Standard and Premium packages. Tell me what you're using and I'll confirm compatibility before we start.
How long does it take?
3 to 7 days depending on the package and complexity. I'll give you a firm timeline once I know the scope.
