Sfoglia categorie
Esplora
Fiverr Pro
Italiano
$
USD
I will build a production ready graphrag chatbot for your docs
Haris Waqar
Informazioni su questo servizio
You're tired of demos that collapse the second you plug in real docs. I'm an AI/ML engineer who builds Retrieval-Augmented Generation (RAG) systems you can actually run: ingestion, embeddings, retrieval, grounding, Fast-API endpoints, and a clear path to deploy on GCP if you want it.
WHAT I NEED FROM YOU BEFORE WE START
- sample documents or access
- tone of voice
- languages
- hosting constraints
- privacy rules
WHAT YOU GET
- A scoped build: [basic/standard/premium bullets match packages]
- Code you can inspect (clean structure, not black box).
- A short handoff: how to ingest new files, tune chunk size/top-k, swap models.
WHAT THIS IS NOT
- Legal/medical certainty or never hallucinates promises.
- Training from scratch mega-models on a gig budget-I'll propose fine-tuning or smaller models only where it fits your scope.
If you paste your use case + document types, I'll reply with what's feasible in each package and realistic timelines-not a sales lecture.
-Haris
Scopri di più su Haris Waqar
Haris Waqar
Ai ML Engineer
- DaPakistan
- Membro daott 2025
- Tempo di risposta medio1 ora
Lingue
Urdu, Inglese
AI/ML Gen Ai Engineer
I specialize in developing, fine-tuning, and deploying AI models for real-world applications. My work focuses on generative AI (like stable diffusion for unique image/video generation), speech recognition (ASR), custom voice bot (Voice-to-Voice) integration and computer vision—bridging advanced research with practical, user-friendly solutions. I combine technical depth in deep learning, NLP, LLMs and CV. Whether you need a creative generative AI tool, a specialized ASR system, or an any AI feature, I deliver efficient, innovative, and production-ready solutions.
FAQ
Do you build with LangChain/LlamaIndex or raw APIs? Which do you prefer?
I would prefer Fast-API integration with LangChain.
Can you use Gemini / OpenAI—who pays tokens?
Yes, I'll use Gemini, as open ai models are not good enough.
My docs are multilingual—does that affect chunking or embeddings?
No, It does not effect chunking of the docs.
Can we host everything on GCP and keep data in our account?
Yes, we can host everything on GCP using containerization and docker depending upon the package you select.
What happens if retrieval returns wrong snippets—how do we improve?
We can solve by these practical order of operations: log failures → check if truth is in top-k → tune chunking + hybrid + filters → add reranker → tune query rewriting → corpus cleanup. That sequence avoids random tuning and usually fixes the most cases per hour spent.
Do you offer post-delivery tweaks or a maintenance window?
Yes, based on the selected package
