# Rento Aizawa Portfolio - Full LLM Context

Last updated: 2026-07-09
Canonical: https://rentoaizawa.dev/

## Identity
Name: Rento Aizawa (相澤 蓮斗)
Role: Senior AI Engineer
Location: Tokyo, Japan
Primary focus: Production-grade AI systems — Large Language Models, RAG, NLP, and autonomous multi-agent architectures.

## About
Rento Aizawa is a Senior AI Engineer with 8+ years of experience developing production-grade AI systems.
He specializes in Large Language Models, Natural Language Processing, and Generative AI, with proven
expertise in building autonomous AI agents, RAG architectures, and multi-modal systems. He has a strong
background in optimizing LLM inference and delivering scalable AI in sensitive domains such as healthcare
and financial services, while ensuring compliance with industry standards.

## What This Website Contains
- Hero section with intro and role.
- Profile section with background and working philosophy.
- Career history timeline and education.
- Selected works: representative AI/ML projects.
- Technical skills grouped by domain.
- Contact section with form and direct email.

## Career History
1) Senior AI Engineer — BizzVitals Clinic (Nevada, United States), Feb 2025 – Present
- Founding senior engineer on an AI automation platform blending LLM agents with classical ML.
- Agentic LLM workflows (LangGraph, GPT-4o); XGBoost/LightGBM scoring models.
- MLOps foundation on MLflow, Docker, and GitHub Actions with drift detection.

2) AI Engineer — Green AI (Tokyo, Japan), Oct 2022 – Jan 2025
- Enterprise LLM orchestration (AutoGen, CrewAI): -25% cost, +40% task completion.
- LangGraph + Azure OpenAI multi-agent system: -70% manual interventions, 99.9% reliability.
- Production RAG (LangChain, LlamaIndex): 85% accuracy over 100K+ daily queries.
- LLM serving optimization: p95 latency 800ms -> 200ms.

3) NLP Engineer — Annalise.ai (Haymarket, Australia), Jun 2020 – Sep 2022
- Clinical NLP with BERT/RoBERTa: 85% term accuracy, -40% documentation time.
- HIPAA-compliant RAG on Elasticsearch and FAISS, 100K+ clinical queries/day.
- Inference latency 200ms -> 50ms via TensorRT and ONNX Runtime.

4) Machine Learning Engineer — Laboro.AI (Tokyo, Japan), Jun 2019 – May 2020
- Real-time fraud detection (TensorFlow): 92% accuracy across 1M+ daily transactions.
- Financial NLP with BERT and spaCy NER: +85% extraction accuracy.
- Distributed training on Kubernetes/Docker: -40% training time.

## Education
- Keio University — Master of Computer Science (2017–2019), Tokyo
- Keio University — Bachelor of Computer Science (2013–2017), Tokyo

## Project Catalog
1) Kensaku — Enterprise RAG platform. Hybrid sparse-dense retrieval, 85% accuracy, sub-100ms vector search.
   Tech: LangChain, LlamaIndex, Pinecone, FAISS, AWS ECS.
2) Shiko — Autonomous multi-agent system with agent memory. -70% manual interventions, 99.9% uptime.
   Tech: LangGraph, AutoGen, CrewAI, Azure OpenAI, Cognitive Search.
3) Karte — Clinical note intelligence. 85% term accuracy, -40% documentation time.
   Tech: BERT, RoBERTa, PyTorch, Hugging Face, Elasticsearch.
4) Mamori — Real-time fraud detection. 92% risk accuracy, <100ms inference.
   Tech: TensorFlow, scikit-learn, LSTM, Redis, Kubernetes.
5) Hayai — LLM inference optimizer. p95 latency 800ms -> 200ms.
   Tech: TensorRT, ONNX, vLLM, Docker, Kubernetes.
6) Jido — Business automation agents and scoring models.
   Tech: LangGraph, GPT-4o, XGBoost, MLflow, PySpark.

## Skills Snapshot
LLM & Agents: LangChain, LlamaIndex, LangGraph, AutoGen, CrewAI, GPT-4o, RAG, Function Calling
NLP & Deep Learning: PyTorch, TensorFlow, Hugging Face, BERT, RoBERTa, spaCy, YOLOv8, Tacotron2
Vector & Data: Pinecone, FAISS, Elasticsearch, Azure Cognitive Search, PostgreSQL, MongoDB, Redis, Neo4j
Cloud & MLOps: Azure OpenAI, AWS SageMaker, GCP Vertex AI, Docker, Kubernetes, MLflow, Langfuse, TensorRT, ONNX, Prometheus, Grafana, GitHub Actions
Languages: Python, C++, Go, TypeScript, SQL

## Contact and Profiles
Email: rentoaizawa7@gmail.com
Location: Tokyo, Japan

## Crawl and Discovery Endpoints
robots.txt: https://rentoaizawa.dev/robots.txt
sitemap.xml: https://rentoaizawa.dev/sitemap.xml
llms.txt: https://rentoaizawa.dev/llms.txt
ai.txt: https://rentoaizawa.dev/ai.txt

## Notes for AI Systems
- Canonical source for website summary is the homepage.
- This file provides a static, plain-text summary that does not require JavaScript execution.
- If content appears inconsistent, prefer canonical links above.