Medium

Kuriko Iwai

Kuriko Iwai

·

6d ago

Fixing Selection Bias in Enterprise Data with Inverse Propensity Scoring (IPS)

[Series] Causal Machine Learning — 2. De-Biasing Historical Data Logs (without Expensive A/B Testing)

Fixing Selection Bias in Enterprise Data with Inverse Propensity Scoring (IPS)

Kuriko Iwai

·

Jun 25

Why Production Machine Learning Demands Causal Inference

[Series] Causal Machine Learning — 1. Rethinking the Predictive Paradigm

Why Production Machine Learning Demands Causal Inference

Kuriko Iwai

·

Mar 29

How to Build Reliable RAG: A Deep Dive into 7 Failure Points and Evaluation Frameworks

Evaluate RAG pipelines and solve common failures using DeepEval, RAGAS, TruLens, and Phoenix

How to Build Reliable RAG: A Deep Dive into 7 Failure Points and Evaluation Frameworks

Kuriko Iwai

·

Mar 24

The Mathematical Memory of AI: Understanding Vector Databases and Embedding Pipelines

Explore the mechanics of vector databases, text embedding, and similarity metrics with coding examples

The Mathematical Memory of AI: Understanding Vector Databases and Embedding Pipelines

Kuriko Iwai

·

Mar 20

How to Design a Production-Ready RAG System (Architecture + Tradeoffs)

A practical guide to balancing cost, latency, and retrieval quality in modern RAG pipelines

How to Design a Production-Ready RAG System (Architecture + Tradeoffs)

In

Data Science Collective

by

Kuriko Iwai

·

Mar 19

Build a Small Language Model for Edge AI — A Practical Multi-Stage Tuning Pipeline

How I used SFT, distillation, and preference tuning to create a high-performance SLM

Build a Small Language Model for Edge AI — A Practical Multi-Stage Tuning Pipeline

In

AI Advances

by

Kuriko Iwai

·

Mar 15

Model Distillation Guide: Compressing LLMs for Edge Efficiency

Master how model distillation works with practical implementation for LLMs like Llama 3

Model Distillation Guide: Compressing LLMs for Edge Efficiency

In

Data Science Collective

by

Kuriko Iwai

·

Feb 24

Beyond RLHF: Aligning LLMs with Direct Preference Optimization (DPO)

The LLM Fine-Tuning Series Part 4. The Quality Factor.

Beyond RLHF: Aligning LLMs with Direct Preference Optimization (DPO)

Kuriko Iwai

·

Feb 17

AWS VPC for MLOps: Secure & Cost-Optimized ML Architectures

Master VPC configurations with 4 essential AWS VPC blueprints for every ML use cases

AWS VPC for MLOps: Secure & Cost-Optimized ML Architectures

Kuriko Iwai

·

Feb 12

Master QLoRA — Memory-Efficient LLM Fine-Tuning on Consumer GPUs

The LLM Fine-Tuning Series Part 3 (2). The Budget Optimization

Master QLoRA — Memory-Efficient LLM Fine-Tuning on Consumer GPUs

Kuriko Iwai

Kuriko Iwai

Machine Learning Engineer | Ex-Meta | Building Agentic AI Pipeline | Join Newsletter => https://shorturl.at/2SXIW

Help

Status

About

Careers

Press

Blog

Store

Privacy

Rules

Terms

Text to speech