Nicolas Méndez Granton 👋

I design and deliver AI systems that don’t just answer questions—they take action in production. With 15+ years in tech and the last 5 focused on AI, I build multi-tenant, cloud platforms that embed Large Language Models (LLMs) into real-world workflows. My work spans agentic systems that complete multi-step transactions to retrieval-augmented assistants serving customers 24/7 across WhatsApp, web, and other channels. Recent skill expansion: model customization (SFT/PEFT/LoRA/QLoRA), RLHF/DPO preference training, and robust eval & safety pipelines. See my work: https://nicolasmendez.com.ar Core strengths - AI Agent Development: Tool-using agents with reliable state & guardrails; deep integrations with CRMs/ERPs/APIs to autonomously resolve requests. - Model Customization & Alignment: Instruction/SFT, PEFT (LoRA/QLoRA), parameter-efficient fine-tunes on Llama/Mistral/GPT-style models; reward modeling + RLHF/DPO. - Evaluation & Safety: Eval harnesses, preference data curation, A/B tests, adversarial tests, safety filters. - Conversational AI: RAG, vector search, multilingual flows, compliance-aware conversations; WhatsApp Cloud API. - Cloud & MLOps: Node.js microservices, async processing, AWS; scalable & secure API design. Highlights - Designed and launched Paula AI, a conversational commerce assistant automating up to 60% of customer interactions for mid-sized retailers. - Built an omnichannel sync platform for large retail brands, ensuring real-time stock, price, and order alignment between physical stores and online channels. - Hands-on with OpenAI GPT family, Claude 3, Llama 3, and open-source stacks; experienced in custom integrations and cost/reliability tuning. Open to Full-time remote roles (AI/LLM Engineer, AI Solutions Architect, Applied LLM/ML Engineer, AI Product Manager) or consulting engagements to integrate/customize LLMs, align models via RLHF/PEFT, and productionize agentic automations. Keywords AI Solutions Architect · AI/LLM Engineer · Applied ML Engineer · AI Product Manager · Conversational AI · Fine-tuning (SFT) · PEFT · LoRA/QLoRA · RLHF · DPO · Reward Modeling · Alignment · Safety/Guardrails · RAG · Prompt Distillation · vLLM · TensorRT-LLM · Hugging Face TGI · LangGraph · LlamaIndex · Vector Databases (Pinecone/Weaviate/pgvector) · API Integration · Microservices · Node.js · AWS · Shopify · Magento · WooCommerce · WhatsApp Cloud API · Omnichannel E-commerce · GPT-4o