Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

AI Agents, RAG Pipelines, Voice AI, LangGraph Workflows, Full Stack AI Apps. I build production-ready agentic systems that handle real users, real workflows, and real business problems. Building an AI agent that works in a demo is easy. Building one that actually runs your business, handles calls, manages tasks, talks to your tools, and keeps working when things get messy, that's what I do. My recent work includes: Building a production multi-agent platform with 15+ specialized agents running 25+ concurrent instances across 7 channels including chat, voice, email, SMS, WhatsApp, and browser, integrated with 20+ third-party services. Developing voice AI agents for inbound and outbound calls with lead qualification, smart routing, and automated post-call summaries, cutting manual call handling time by over 60%. Building RAG systems using semantic search and hybrid retrieval across short and long-term memory indexes. The result is an AI that pulls accurate answers from a business's own documents and data, reducing hallucinations by up to 40% and improving response relevance scores significantly compared to standard LLM setups. Shipping full stack AI apps end to end, agent logic, Python backend, and React or Next.js frontend, so clients get one working product instead of three disconnected pieces. I don't just connect APIs. I think about what the system needs to do six months from now. That's the difference between a demo and something you can run your business on. What I build: →Multi-agent systems with LangGraph and OpenAI Agents SDK, stateful, memory-enabled, and handling complex decision logic. →RAG pipelines with semantic search, hybrid retrieval, and persistent memory across short and long-term indexes. →Voice AI agents with real-time transcription, call routing, and post-call summaries. →Full stack AI apps with FastAPI backends and React or Next.js frontends. →LLM integrations across OpenAI, Claude, and Gemini with provider-agnostic routing. →Browser automation agents that run multi-step tasks autonomously. Tech stack: Python, FastAPI, LangGraph, LangChain, LangMem, OpenAI Agents, Hugging Face, Pinecone, FAISS, ChromaDB, Redis, MongoDB, PostgreSQL, Twilio, ElevenLabs, Node.js, React, Next.js, Docker, GitHub Actions, AWS, Google Cloud, MCP, WebSockets. Have a project? Message me with what you're trying to build and where it's getting stuck. I'll come back with a concrete plan, no charge.…AI Agents, RAG Pipelines, Voice AI, LangGraph Workflows, Full Stack AI Apps. I build production-ready agentic systems that handle real users, real workflows, and real business problems. Building an AI agent that works in a demo is easy. Building one that actually runs your business, handles calls, manages tasks, talks to your tools, and keeps working when things get messy, that's what I do. My recent work includes: Building a production multi-agent platform with 15+ specialized agents running 25+ concurrent instances across 7 channels including chat, voice, email, SMS, WhatsApp, and browser, integrated with 20+ third-party services. Developing voice AI agents for inbound and outbound calls with lead qualification, smart routing, and automated post-call summaries, cutting manual call handling time by over 60%. Building RAG systems using semantic search and hybrid retrieval across short and long-term memory indexes. The result is an AI that pulls accurate answers from a business's own documents and data, reducing hallucinations by up to 40% and improving response relevance scores significantly compared to standard LLM setups. Shipping full stack AI apps end to end, agent logic, Python backend, and React or Next.js frontend, so clients get one working product instead of three disconnected pieces. I don't just connect APIs. I think about what the system needs to do six months from now. That's the difference between a demo and something you can run your business on. What I build: →Multi-agent systems with LangGraph and OpenAI Agents SDK, stateful, memory-enabled, and handling complex decision logic. →RAG pipelines with semantic search, hybrid retrieval, and persistent memory across short and long-term indexes. →Voice AI agents with real-time transcription, call routing, and post-call summaries. →Full stack AI apps with FastAPI backends and React or Next.js frontends. →LLM integrations across OpenAI, Claude, and Gemini with provider-agnostic routing. →Browser automation agents that run multi-step tasks autonomously. Tech stack: Python, FastAPI, LangGraph, LangChain, LangMem, OpenAI Agents, Hugging Face, Pinecone, FAISS, ChromaDB, Redis, MongoDB, PostgreSQL, Twilio, ElevenLabs, Node.js, React, Next.js, Docker, GitHub Actions, AWS, Google Cloud, MCP, WebSockets. Have a project? Message me with what you're trying to build and where it's getting stuck. I'll come back with a concrete plan, no charge.

Muhammad Dayyan





AI Agents, RAG Pipelines, Voice AI, LangGraph Workflows, Full Stack AI Apps. I build production-ready agentic systems that handle real users, real workflows, and real business problems. Building an AI agent that works in a demo is easy. Building one that actually runs your business, handles calls, manages tasks, talks to your tools, and keeps working when things get messy, that's what I do. My recent work includes: Building a production multi-agent platform with 15+ specialized agents running 25+ concurrent instances across 7 channels including chat, voice, email, SMS, WhatsApp, and browser, integrated with 20+ third-party services. Developing voice AI agents for inbound and outbound calls with lead qualification, smart routing, and automated post-call summaries, cutting manual call handling time by over 60%. Building RAG systems using semantic search and hybrid retrieval across short and long-term memory indexes. The result is an AI that pulls accurate answers from a business's own documents and data, reducing hallucinations by up to 40% and improving response relevance scores significantly compared to standard LLM setups. Shipping full stack AI apps end to end, agent logic, Python backend, and React or Next.js frontend, so clients get one working product instead of three disconnected pieces. I don't just connect APIs. I think about what the system needs to do six months from now. That's the difference between a demo and something you can run your business on. What I build: →Multi-agent systems with LangGraph and OpenAI Agents SDK, stateful, memory-enabled, and handling complex decision logic. →RAG pipelines with semantic search, hybrid retrieval, and persistent memory across short and long-term indexes. →Voice AI agents with real-time transcription, call routing, and post-call summaries. →Full stack AI apps with FastAPI backends and React or Next.js frontends. →LLM integrations across OpenAI, Claude, and Gemini with provider-agnostic routing. →Browser automation agents that run multi-step tasks autonomously. Tech stack: Python, FastAPI, LangGraph, LangChain, LangMem, OpenAI Agents, Hugging Face, Pinecone, FAISS, ChromaDB, Redis, MongoDB, PostgreSQL, Twilio, ElevenLabs, Node.js, React, Next.js, Docker, GitHub Actions, AWS, Google Cloud, MCP, WebSockets. Have a project? Message me with what you're trying to build and where it's getting stuck. I'll come back with a concrete plan, no charge.…AI Agents, RAG Pipelines, Voice AI, LangGraph Workflows, Full Stack AI Apps. I build production-ready agentic systems that handle real users, real workflows, and real business problems. Building an AI agent that works in a demo is easy. Building one that actually runs your business, handles calls, manages tasks, talks to your tools, and keeps working when things get messy, that's what I do. My recent work includes: Building a production multi-agent platform with 15+ specialized agents running 25+ concurrent instances across 7 channels including chat, voice, email, SMS, WhatsApp, and browser, integrated with 20+ third-party services. Developing voice AI agents for inbound and outbound calls with lead qualification, smart routing, and automated post-call summaries, cutting manual call handling time by over 60%. Building RAG systems using semantic search and hybrid retrieval across short and long-term memory indexes. The result is an AI that pulls accurate answers from a business's own documents and data, reducing hallucinations by up to 40% and improving response relevance scores significantly compared to standard LLM setups. Shipping full stack AI apps end to end, agent logic, Python backend, and React or Next.js frontend, so clients get one working product instead of three disconnected pieces. I don't just connect APIs. I think about what the system needs to do six months from now. That's the difference between a demo and something you can run your business on. What I build: →Multi-agent systems with LangGraph and OpenAI Agents SDK, stateful, memory-enabled, and handling complex decision logic. →RAG pipelines with semantic search, hybrid retrieval, and persistent memory across short and long-term indexes. →Voice AI agents with real-time transcription, call routing, and post-call summaries. →Full stack AI apps with FastAPI backends and React or Next.js frontends. →LLM integrations across OpenAI, Claude, and Gemini with provider-agnostic routing. →Browser automation agents that run multi-step tasks autonomously. Tech stack: Python, FastAPI, LangGraph, LangChain, LangMem, OpenAI Agents, Hugging Face, Pinecone, FAISS, ChromaDB, Redis, MongoDB, PostgreSQL, Twilio, ElevenLabs, Node.js, React, Next.js, Docker, GitHub Actions, AWS, Google Cloud, MCP, WebSockets. Have a project? Message me with what you're trying to build and where it's getting stuck. I'll come back with a concrete plan, no charge.

Available to hire

AI Agents, RAG Pipelines, Voice AI, LangGraph Workflows, Full Stack AI Apps. I build production-ready agentic systems that handle real users, real workflows, and real business problems.

Building an AI agent that works in a demo is easy. Building one that actually runs your business, handles calls, manages tasks, talks to your tools, and keeps working when things get messy, that’s what I do.

My recent work includes:

Building a production multi-agent platform with 15+ specialized agents running 25+ concurrent instances across 7 channels including chat, voice, email, SMS, WhatsApp, and browser, integrated with 20+ third-party services.

Developing voice AI agents for inbound and outbound calls with lead qualification, smart routing, and automated post-call summaries, cutting manual call handling time by over 60%.

Building RAG systems using semantic search and hybrid retrieval across short and long-term memory indexes. The result is an AI that pulls accurate answers from a business’s own documents and data, reducing hallucinations by up to 40% and improving response relevance scores significantly compared to standard LLM setups.

Shipping full stack AI apps end to end, agent logic, Python backend, and React or Next.js frontend, so clients get one working product instead of three disconnected pieces.

I don’t just connect APIs. I think about what the system needs to do six months from now. That’s the difference between a demo and something you can run your business on.

What I build:
→Multi-agent systems with LangGraph and OpenAI Agents SDK, stateful, memory-enabled, and handling complex decision logic.
→RAG pipelines with semantic search, hybrid retrieval, and persistent memory across short and long-term indexes.
→Voice AI agents with real-time transcription, call routing, and post-call summaries.
→Full stack AI apps with FastAPI backends and React or Next.js frontends.
→LLM integrations across OpenAI, Claude, and Gemini with provider-agnostic routing.
→Browser automation agents that run multi-step tasks autonomously.

Tech stack: Python, FastAPI, LangGraph, LangChain, LangMem, OpenAI Agents, Hugging Face, Pinecone, FAISS, ChromaDB, Redis, MongoDB, PostgreSQL, Twilio, ElevenLabs, Node.js, React, Next.js, Docker, GitHub Actions, AWS, Google Cloud, MCP, WebSockets.

Have a project? Message me with what you’re trying to build and where it’s getting stuck. I’ll come back with a concrete plan, no charge.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Language

English

Fluent

Work Experience

Lead AI Engineer at Growth Rune

May 1, 2025 - Present

Led agent orchestration and multi-agent architecture for a production autonomous agent platform powering humanlike AI teammates and agentic avatars across 7 communication channels. Architected 15+ specialized sub-agents with autonomous decision making and inter-agent coordination, using Redis as the distributed state store and RQ-based queues. Managed 100+ agent tools (messaging, browser automation, social graph management, image/video/3D content generation, form building, web scraping, real-time notifications) with auditable per-tool capabilities. Implemented a provider-agnostic LLM infrastructure enabling dynamic routing to models from any provider, streaming token delivery, and zero-code model switching. Built hybrid memory and RAG with per-user memory persistence across chat, voice, video, SMS, and email to preserve context. Developed real-time systems including WebSocket-first streaming, live agent observation, interactive control handoffs between agent and human, voice/video call

AI Engineer at Octalop Technologies

December 1, 2024 - April 1, 2025

Built and deployed AI-powered products including LLM-driven conversational agents, real-time multilingual voice bots, and computer vision pipelines using FastAPI, LangChain, and OpenAI in production environments. Implemented agent architectures for autonomous task execution, context-aware RAG systems for knowledge retrieval, and domain-adapted LLMs via LoRA fine-tuning to reduce compute while improving task-specific accuracy.

Generative AI Intern at DevSoft

July 1, 2024 - November 1, 2024

Developed RAG pipelines with Hugging Face Transformers for knowledge-augmented chatbots; performed diffusion model fine-tuning with LoRA for text-to-image generation with reduced compute overhead.