For Decision Makers
The signal beneath the noise — value, impact, and what it means for your next move. Two minutes.
IT PM + AI/ML Engineer
The interface between human intent and machine intelligence
I build the systems where human judgment meets artificial cognition — practical LLM architectures that know their own limits.
For Those Who Look Deeper
Descend into the architecture — constraints, tradeoffs, and the implementation patterns where novelty crystallizes into working systems.
JD Concierge
Paste a job description to get an evidence-grounded fit analysis.
0 / 10000 Remaining: 10000
Analyzing...
Analysis based on provided JD text and publicly documented portfolio evidence.
Skills Demonstrated
Prompt Engineering
RAG
Agentic Workflows
Change Enablement
Python/Automation
Evaluation
Four Investigations
Each project maps a different territory where language models meet the resistance of the real world.
Automated statement-to-receipt reconciliation for faster, cleaner tax submission workflows.
- Reduced manual receipt matching and exception triage effort.
- Improved categorization consistency for tax-ready reporting.
- Added ambiguity handling for difficult vendor names.
AI-powered Japanese tax analysis built on official NTA and e-Gov sources, with verifiable citations for every answer.
- Trusted answers for Japanese tax law based on official government data.
- {"Four integrated solutions"=>"Tax AI Advisor, Regulatory Monitoring, Legal Research and Comparison, and Multi-Tenant Platform."}
- Citation-first response design focused on verifiability over AI guesswork.
Produced grounded Japanese/English ceremony scripts with strong style control and iterative refinement.
- Used 10 prior scripts as grounding corpus for consistent quality.
- Generated theme-faithful bilingual output for an Ancient Grecian concept.
- Reduced rewriting time with collaborative iteration cycles.
Automated reporting and incident intelligence to improve executive visibility and decision making.
- Automated PM reporting and executive dashboard workflows.
- Unified insights across Confluence, Jira, SharePoint, Power Automate, and Power BI.
- Added coding-agent parsers for hundreds of MB of incident data trend analysis.
Now / Next
About / Contact
Agents, prompts, retrieval architectures, and automation woven together — because the most interesting outcomes emerge at the intersection of structured intent and machine intuition.