I will build an agentic ai workflow with langchain, n8n, llama 3 and voice ai
AI Engineer
About this Gig
I work with the most capable AI models and automation tools available today. For LLM tasks I use LLaMA 3, Qwen 3, Mistral, GPT-4.5, Claude Sonnet 4.6, and Gemini 2.0 selecting the right model based on your use case and budget. For inference speed, LLaMA 3 and Mixtral run through Groq AI for ultra-low latency in real-time agent applications.
For voice AI I build full pipelines using OpenAI Whisper for multilingual speech-to-text, NVIDIA Parakeet TDT for real-time streaming ASR, and ElevenLabs for voice cloning and natural text-to-speech synthesis.
My automation stack runs on LangGraph for stateful multi-agent orchestration, LangChain for RAG pipelines and tool-calling, and n8n for no-code visual workflow automation. Supporting libraries include Hugging Face Transformers, PyTorch, spaCy, FAISS, Pinecone, and LlamaIndex.
On the platform side I integrate with Shopify, WooCommerce, PrestaShop, and Magento for ecommerce automation, Gmail and Google Workspace for productivity workflows, and Facebook, Instagram, and Google Ads for social media automation all connected through REST APIs and n8n pipelines.
My Portfolio
FAQ
Which AI models do you specialize in fine-tuning?
I specialize in fine-tuning and deploying a wide range of modern foundation models: Open-source LLMs: Qwen 3 (Alibaba MoE & dense), LLaMA 3, Mistral / Mixtral, Falcon, BERT, GPT-2 — using LoRA, QLoRA, PEFT, and RLHF techniques via Hugging Face Transformers. Proprietary APIs: OpenAI GPT-4.5, Google
Can you integrate voice AI into my existing application?
Yes. I build complete voice AI pipelines that integrate directly into web, mobile, or backend applications. This includes: ASR (speech-to-text): Whisper API for high-accuracy multilingual transcription, or Parakeet TDT via NVIDIA NeMo for low-latency real-time streaming ASR. TTS (text-to-speech):
Which models do you work with?
Model choice depends on your task, budget, and deployment needs: Claude Sonnet 4.6 — Best for agentic workflows, long-document reasoning, and safe, instruction-following chatbots. GPT-4.5 — Ideal for RAG pipelines, function calling, and general-purpose enterprise applications. Gemini 2.0 — Best f
