Browse categories
Explore
Fiverr Pro
English
$
USD
I will build langchain or langgraph multi agent ai pipelines
MalakKhan
About this gig
You want AI agents that work in production not just in a tutorial notebook.
I've shipped multi-agent systems at a real biotech company on GCP. I know what breaks: agents looping on bad state, rate limits killing jobs silently, JSON truncation mid-response. I've fixed all of it in live systems.
What I build:
LangChain chains with tool use, memory, structured output
LangGraph state machines for complex branching workflows
RAG pipelines let agents query your own documents or database
Rate-limit handling and circuit breaker patterns
Prompt optimization for consistent, reliable outputs
FastAPI or Streamlit frontend (Premium tier)
Real projects I've shipped:
Agentic AI system deployed live on GCP (InTime BioTech)
Patched LLM rate-limit circuit breaker in production
Multi-agent prompt chains optimized to fix logic bottlenecks
This gig is for you if:
You want to automate a multi-step workflow with LLMs
You need a document Q&A system (RAG) for internal data
Your existing agent pipeline is breaking or behaving unpredictably
Message me before ordering I'll scope it honestly.
Get to know MalakKhan
MalakKhan
- FromPakistan
- Member sinceMay 2024
- Avg. response time1 hour
Languages
English
Hey, I'm Malak — an AI/ML engineer who builds things that actually work in production, not just in notebooks. I've deployed agentic AI systems on GCP, built multi-agent LangGraph pipelines, and shipped full-stack AI apps with FastAPI and Streamlit. My work spans RAG pipelines, NLP with transformer models, real-time ML APIs, and Gemini Vision integrations. If you need an AI feature built right the first time — or need someone to untangle a system that's misbehaving — let's talk. I'm quick to respond, clear about scope, and I won't overpromise.
My Portfolio
FAQ
What's the difference between LangChain and LangGraph?
LangChain is best for linear chains and simple tool use. LangGraph handles complex workflows where agents loop, branch, or pass state between each other. I'll recommend the right one for your use case — just message me first and I'll tell you within minutes.
Can you fix my existing broken agent pipeline?
Yes. Debugging broken pipelines is something I'm specifically experienced in — rate limit failures, agent loops, truncation issues, and state conflicts. Share your code and describe what's going wrong before ordering and I'll give you an honest assessment.
Which LLM providers do you support?
OpenAI (GPT-4o), Google Gemini, Anthropic Claude, and open-source models via Groq or Ollama. I work with your existing API keys and setup. If you're unsure which model fits your use case, I'll recommend one based on your budget and requirements.
Can you build a RAG system for my company's documents?
Yes — RAG with Pinecone, Chroma, or Weaviate is included in the Premium package and can be added to Standard as an upgrade. Supported formats: PDF, DOCX, Notion exports, and website content. Message me with your document type and I'll confirm the right approach.
Will the agent work reliably, not just in testing?
I build in error handling, fallback logic, and rate limit management from the start — the same patterns used in my production deployments. I also include logging so you can monitor agent behavior after delivery. Reliability is built in, not added later.
Do I own the code after delivery?
Yes. Full source code is delivered with every package. You own it completely and can extend or modify it however you like after delivery. No restrictions.
