Sfoglia categorie
Esplora
Fiverr Pro
Italiano
$
USD
Sembra che questo servizio sia in sospeso
I will build agentic rag, knowledge graph and multi agent ai systems
Adnan
Informazioni su questo servizio
Production-grade agentic RAG and multi-agent AI that holds up in real use, not just demos. I build agent systems with LangGraph or CrewAI, or I configure and deploy the pre-built Strands Harness runtime for you.
What I Build:
- Agentic RAG: Grounded retrieval over Vector DBs, knowledge graphs and hybrid search, with source citations on every answer.
- Multi-Agent Orchestration: Task routing, delegated subagents, and persistent memory that survives restarts.
- Strands Harness (Premium): The pre-built agent runtime, configured and deployed by me, not rebuilt from scratch. Session persistence, automatic context management, background task execution and approval gates on every tool call, on AWS or Docker.
- Model Agnostic: Runs on OpenAI, Anthropic, Google or Amazon Bedrock, so you can switch providers without a rewrite.
- Guardrails: Human approval and authorization policies so no tool call runs unchecked.
- Add-ons: PDF table extraction, metadata filtering, Neo4j graphs, read-only database agents.
You own the code and the configuration. You get tested source code, architecture documentation and deployment support. Message me for a technical consultation before you order.
Scopri di più su Adnan
Adnan
Building AI systems and the Python backends that serve them
- DaBangladesh
- Membro daapr 2025
Lingue
Inglese
Hi, I'm Adnan. I build AI systems and the Python backends that serve them.
My work covers four areas: agentic RAG and multi-agent systems that cite their sources; AI chat widgets for your website; Python FastAPIs with a real database and tests; and third-party API integrations that handle webhooks and retries.
I work in Python throughout, with FastAPI, PostgreSQL, Redis, LangGraph, and vector databases.
Thank you for reading. I would be glad to hear about your project and advise on the most suitable approach.
Portfolio: adnanmd[dot]pythonanywhere[dot]com
Il mio portfolio
FAQ
What is the difference between standard RAG and Agentic RAG?
Standard RAG retrieves static context for prompt injection. Agentic RAG uses autonomous agents with decision logic, tool calling, and iterative retrieval for multi-step reasoning.
Which vector databases and frameworks do you support?
LangGraph, LangChain, Pinecone, ChromaDB, Qdrant and PGVector, integrated with OpenAI, Claude, DeepSeek or open-source LLaMA models.
Can I keep control of what my agent is allowed to do?
Yes. Every tool call passes through an intervention gate first — full human approval, AI risk scoring, plain-language rules, or formal authorization policies that subagents inherit automatically. Delegates cannot bypass your rules.
Can you work with my PDFs, scanned documents and spreadsheets?
Yes. I extract tables, charts, figures and scanned pages into clean structured data, with page-level tracking so you can trace any figure back to its source page.
Can the agent answer questions against my own database?
Yes, as an add-on. Your team asks in plain English and gets an answer with the SQL shown, so they can check the logic rather than trust a black box. It runs on a **read-only** database connection with a statement allowlist and a row cap — the agent cannot alter your data.
What exactly is the "agent harness" in the Premium package — are you building me a framework?
No. Strands Harness is an existing, maintained runtime and I configure and deploy it for you rather than building or forking anything. You get sessions, long-term memory, context management, subagents, background tasks and approval gates wired to your data and tools.
Do I get long-term memory and state persistence?
Yes — session state, persistent vector storage and conversation memory are included in Standard and Premium.
