Security ResearcherRed Team Expert

KenzieTandera

I build systems end to end — then I try to break them. Full-stack engineering with a security researcher’s instinct for where things give way.

Based in
Osaka, Japan
Studying at
HAL Osaka Vocational College
02

About

I’m a student at HAL Osaka, building production software by day and taking it apart by night.

Based in
Osaka, Japan
Studying
HAL Osaka Vocational College
Mentored by
Ex-Netflix Senior Cloud Engineer
Focus
Full-stack systems & offensive security

Most of what I know came from shipping things that had to actually work — APIs that stay up, interfaces people can move through without thinking, and tooling that removes the boring parts of a workflow. I write across the stack because the interesting problems rarely respect the boundary between front and back.

The security side started as curiosity and turned into discipline. Working through authorized assessments taught me to read a system the way an attacker does: assume nothing, verify everything, and pay attention to the seams where two components agree to trust each other.

I’m mentored by a former Netflix Senior Cloud Engineer, which reset my sense of what “done” means. Reliability at scale, observability, blast-radius thinking — the habits that separate a project from a product.

03

Experience

NowAugust 2026 — Present

Millelith

  • Co-Founder
  • CTO
  • Security Researcher
  • Software Engineer

Co-founded Millelith in August 2026. As CTO I own the engineering direction, as security researcher I make sure what we build holds up against the people trying to break it, and as software engineer I still build hands-on.

  • Engineering leadership
  • Architecture
  • Security research
  • Software engineering
04

Security research

Authorized.
Verified.
Reported.

Since August 2026, Millelith engagements have taken me through SMEs and household-name enterprises across Indonesia and Japan. The interesting parts stay confidential. The numbers don’t.

Millelith engagementsAug 2026 —

Vulnerabilities reported

149+FindingsValidated & reported

Authorized assessments for banks, a telecom operator, venture capital firms, technology startups, universities, government bodies, and hospitals, from SME to enterprise scale. Every finding reproduced, rated, and handed to the owner with remediation guidance.

Channel
Private programs & direct engagements
Status
Reported to owners
SeverityCritical

Rated critical

50+CVSS 9.0+Critical severity

More than a third scored 9.0 or higher under CVSS v3.1: the class of issue that ends in full compromise if left alone. Reported through coordinated channels under written authorization; the details stay with the owners.

Standard
CVSS v3.1
Handling
Coordinated disclosure
Where the work happenedBankingTelecommunicationsTrading & conglomeratesVenture capitalTech startupsGamingUniversities & collegesGovernmentHealthcareIndonesia & Japan

Selected findings

  1. 01

    Cloud credentials in a public bundle

    An AWS IAM access key cached in the public frontend bundle of a venture capital firm. Reported privately; the key was rotated and the exposure closed.

    Fixed
  2. 02

    Database ports open to the internet

    Publicly reachable database ports across venture capital, university, government, and hospital systems. Reported to each owner; access was restricted to internal networks.

    Fixed
  3. 03

    Leaked JWT signing secret

    A JWT signing secret exposed by a fintech API provider, enough to forge valid session tokens. Reported privately; the secret was rotated.

    Fixed
  4. 04

    Live Azure AD credentials in public JavaScript

    Active Azure AD application credentials shipped inside the public JavaScript of a national telecom operator. Reported privately; the credentials were revoked.

    Fixed
  5. 05

    End-of-life software with public CVEs

    A major Japanese technology and telecom group ran internet-facing software well past end-of-life, exposed to several publicly known CVEs. Reported privately with upgrade paths.

    Reported
  6. 06

    Outdated components, known CVEs

    A global video game publisher shipped production services on end-of-life components carrying published CVEs. Reported privately with remediation guidance.

    Reported
05

Selected work

01AI CV Scorer2025

Matchingg

Scores a résumé against a job description and explains the gap.

Matchingg reads a CV and a job posting, then returns a match score with the reasoning attached — which requirements are covered, which are missing, and what phrasing is costing the candidate points.

The hard part was making the output trustworthy rather than merely confident: structured extraction first, scoring second, so every number traces back to something actually present in the document.

  • TypeScript
  • NestJS
  • PostgreSQL
  • LLM APIs
  • Angular
Matchingg interface showing a CV match score

02CLI Tool2025

Gitauto

Removes the ritual from the commit-and-push loop.

A command-line tool for the repetitive half of version control — staging, conventional commit messages, branch hygiene, and push, collapsed into a single intentional command.

Written to be predictable above all: it never rewrites history you did not ask it to rewrite, and every destructive path requires confirmation.

  • Go
  • Bash
  • Git Plumbing
Gitauto running in a terminal

03Editorial Web2024

RADWIMPS Fansite

A discography and archive built as a love letter.

A fan-built archive for RADWIMPS — releases, timelines, and lyric context, laid out with the kind of typographic care the music deserves.

Entirely front-end craft: responsive editorial layout, image performance discipline, and motion that supports the reading rather than interrupting it.

  • Angular
  • TypeScript
  • CSS
  • GSAP
RADWIMPS fansite discography page

04Security Platform2026

Millelith Internal Tools

Recon, findings, and AI-written reports for authorized assessments. Private.

The internal platform behind Millelith engagements: a NestJS backend and scanner service on Node 20, an Angular 21 console with a dark terminal feel, and a shared TypeScript contracts library that keeps finding schemas, DTOs, and bus messages honest across services.

The recon engine is a Bash pipeline over nmap, ffuf, gobuster, subfinder, dnsx, dig, and whois, with Python helpers for parsing. Findings flow through RabbitMQ into MariaDB via TypeORM; Claude writes the report prose against Zod-validated schemas and Puppeteer renders the PDF. Everything ships as Docker Compose behind Tailscale, tested with node:test and GitHub Actions.

  • TypeScript
  • NestJS
  • Angular
  • TypeORM
  • MariaDB
  • RabbitMQ
  • Claude API
  • Docker
  • Bash

Private repository — no public source or demo.

06

Skills

Security

Offensive testing & research
  • Burp Suite
  • ZAP
  • Nuclei
  • OWASP Top 10
  • API Top 10
  • SANS Top 25
  • HackerOne

Recon

Toolchain in the scanner image
  • Nmap
  • ffuf
  • Gobuster
  • Subfinder
  • dnsx
  • dig
  • whois
  • jq

Exploitation

Kali Linux toolset
  • Metasploit
  • sqlmap
  • Hydra
  • John the Ripper
  • Hashcat
  • Nikto
  • Wireshark
  • Kali Linux

Languages

Daily drivers
  • TypeScript
  • JavaScript
  • Go
  • Python
  • Kotlin
  • Bash
  • PowerShell

Frontend

Interface & motion
  • Angular
  • React
  • Three.js
  • GSAP
  • HTML
  • CSS

Backend

Services & data
  • Node.js
  • NestJS
  • Fastify
  • PostgreSQL
  • MariaDB
  • Redis
  • TypeORM
  • RabbitMQ
  • JWT
  • Puppeteer
  • Claude API
  • Zod

Platform

Ship & operate
  • Docker
  • Tailscale
  • GitHub Actions
  • Vercel
  • DigitalOcean
  • Git
07

Contact

Let’s build
or secure
something.

Open to security engagements under written authorization, partnerships, and problems that don’t fit neatly in either box. Osaka hours, worldwide reach.

© 2026 Kenzie TanderaBuilt with Angular & GSAP