AI/LLM Engineer · NAACL 2025 Published Researcher

Computer Science graduate from UTEC building production AI at the intersection of LLMs, RAG, and agentic systems. I co-authored a paper accepted at NAACL 2025 (Industry Track) on LLM fine-tuning for text generation, and I enjoy taking AI features end-to-end — from prototyping to reliable, scalable deployment on AWS and GCP.

📍 Peru · 🌎 Open to remote AI/LLM Engineering roles 🔗 LinkedIn · ✉️ luis.gutierrez.g@utec.edu.pe

• LLM applications — building with LLM APIs (OpenAI, Anthropic, Gemini) and fine-tuning open models (Llama 2, Mistral, StructLM)
• RAG systems — retrieval, chunking, embeddings, re-ranking and evaluation
• Agentic & multimodal workflows — LLM + computer vision pipelines, tool use, routing and orchestration
• Backend & cloud — Python/Node.js REST APIs, webhooks, JWT auth, Docker, AWS & GCP serverless (Lambda / Cloud Run)

• SkillLab — Platanus Hack 2025 — Sports-clip analysis app: splits 30s video into chunks, extracts keyframes and runs per-chunk LLM + computer-vision analysis to surface technical/tactical insights. Python (FastAPI) backend with Dockerized AWS handlers (analyzer/splitter), React + Vite frontend.
• eC-Tab2Text — Aspect-based text generation from e-commerce product tables. Fine-tuning of Llama 2, Mistral and StructLM with Table2Text metrics plus correctness, faithfulness and fluency evaluation. 📄 Published at NAACL 2025 (Industry Track).
• Invoice Automation — Automates ingestion of PDF/XML invoices from email via webhook, extracts & classifies fields, and exports grouped archives. Node.js + TypeScript REST API on GCP Cloud Run + Firestore.
• Phonance 📱 — Personal-finance mobile app live on the Google Play Store. Flutter app with automatic transaction capture; serverless backend on GCP (Cloud Run + Firestore + API Gateway) with JWT auth (originally deployed on AWS Amplify, later migrated to GCP).
• 7H2 Hackathon Winner 🏆 — Real-time VALORANT "virtual coach" agent that analyzes gameplay frame-by-frame with Gemini and returns round-by-round, actionable feedback.
• Hackathon RIMAC 2024 Winner (Data Challenge) 🏆 — Data-driven solution for well-being, built and pivoted under a tight deadline.

Languages: Python · TypeScript · SQL · Dart AI/LLM: LLM APIs (OpenAI, Anthropic, Gemini) · LLM fine-tuning · RAG · LangChain · embeddings · prompt engineering · LLM + computer vision Cloud & MLOps: AWS (Lambda, Amplify) · GCP (Cloud Run, Firestore, API Gateway) · Azure (Data Factory, Databricks, Data Lake) · Docker · Git Backend & Web: FastAPI · Node.js · REST APIs · webhooks · JWT · React · Next.js · Vue.js

• Software Engineer @ Belcorp
• Research Assistant @ University of Alberta — data ingestion & cleaning pipelines for LLM training
• Data Engineer Intern @ Pacífico Seguros — Azure Data Factory, Databricks, Power BI
• ML Engineer Intern @ Amber — model deployment & MLOps