Open to opportunities

Hikaru M

@hikarum

AI Engineer and Data Science Manager building deployed LLM and analytics solutions that deliver measurable business impact.

Germany

Message

What I'm looking for

I’m looking to build and scale LLM and data products end-to-end—RAG to deployment—using reliable evaluation, CI/CD, and measurable business impact. I want teams that value experimentation, strong engineering practices, and cross-functional collaboration.

I’m an AI Engineer and Data Science Manager with 10+ years of experience building and deploying data and AI solutions across insurance and accounting. I bridge domain experts and engineering teams to turn ambiguous business problems into scalable AI solutions with clear, operational value.

At KPMG, I led the end-to-end development and global rollout of a LLM-powered product, moving from PoC to production in 6 months and delivering ~10,000 hours of annual operational efficiency gains. I built modular LLM pipelines (RAG, reasoning, validation), used OpenAI APIs with tool calling and orchestration frameworks (LangChain, LangGraph), and improved reliability with CI/CD and unit testing of LLM outputs using Robot Framework.

Previously, I drove fraud and audit analytics end-to-end, including metric definition, preprocessing, modeling, ratio analysis, and visualization for multiple internal clients, and I supported audit and ESG assurance after obtaining the USCPA license. I’m also hands-on beyond core work—whether it’s building dashboards, risk reporting, or a web-based price comparison tool using web scraping pipelines.

Experience

Work history, roles, and key accomplishments

Current

Data Science Manager

Current

KPMG Global Solution Group

Nov 2023 - Present (2 years 7 months)

Led end-to-end development and global rollout of an LLM-powered product, quantifying causal business impact through pilots and delivering ~10,000 hours of annual operational efficiency gains. Managed a team of six data scientists and built modular LLM pipelines (RAG, reasoning, validation) with CI/CD optimizations to reduce latency and cost.

LLM Application Development Python OpenAI API LangGraph Azure AI Foundry CI CD Robot Framework Neo4j

Senior Data Scientist

KPMG AZSA LLC

Jul 2021 - Nov 2023 (2 years 4 months)

Drove end-to-end analytics for fraud and audit use cases, defining fraud metrics, performing data preprocessing and modeling, and delivering ratio analysis and visualizations under tight audit deadlines across 5+ internal clients. Designed a self-serve pipeline to monitor monthly/quarterly metrics with clear standards for usability and impact.

Fraud Analytics Audit Analytics R Python SQL Alteryx data preprocessing Modeling Tableau Power BI

Crowdfunding Lead

Bgrass Co., Ltd.

Jun 2022 - Mar 2023 (9 months)

Initiated and led grassroots fundraising for an early-stage startup initiative, independently planning events, conducting outreach via Twitter/X, and building supporter relationships. Raised approximately ¥1M within 3 months despite no prior fundraising infrastructure.

Fundraising Event Planning Outreach Via Twitter X Stakeholder Management Community Building Marketing Communications Leadership

Junior Data Scientist

KPMG AZSA LLC

Jun 2019 - Jun 2021 (2 years)

Built a consolidated financial metrics dashboard for a client proposal within 3 months using Tableau and Alteryx, contributing to winning a competitive bidding engagement. Initiated and led monthly innovation knowledge-sharing to improve collaboration and idea exchange across teams.

Tableau Alteryx SQL Financial Metrics Data Visualization Dashboarding Python

Risk Analyst

Tokio Marine & Nichido Fire Insurance Co., Ltd.

Apr 2016 - Mar 2019 (2 years 11 months)

Performed quarterly catastrophe-risk reporting with in-depth portfolio analysis and delivered risk insights to accounting, reinsurance, and external stakeholders. Produced recurring analyses to support risk understanding and decision-making.

Portfolio Analysis Risk & Analytics SAS Data Analysis