Natalie Phillips
Fraud & Risk Analyst · Investigator · Independent Builder
My professional background is in fraud investigation, risk and complex operations, with experience analysing suspicious activity, identifying anomalies, assessing financial exposure and making evidence-led decisions.
Alongside that, I build independently using AI-assisted development — from consumer products and APIs to operational tools and automated systems.
Investigate
Follow the evidence. Patterns, exceptions, inconsistencies and the thing that doesn't quite make sense.
Define
Make the problem concrete. What are the rules? What does good look like? What can go wrong?
Build
When a tool would help, build one. Start with the specification rather than the interface.
Verify
Question the result. Test assumptions. Check failure states. Look for false confidence.
Different kinds of work, same standard: evidence first, rules explicit, outputs checked.
- 3.5 years
Investigations · Refunds
Complex operations
- Fraud & risk
Pattern analysis · Evidence
Financial judgement
- Shipped product
MELAIRE · iOS
A consumer app
- Independent building
Apps · APIs · Automation
Dashboards
Complex operations
Revenue Protection investigations
Independent building
2022
2023
2025
- 2022
Complex operations
- 2023
Revenue Protection investigations
- 2025
Independent building
01 — Two sides of my work
Two sides of my work
My professional experience and my independent projects are related by how I think, but they are not the same thing. I want both to be visible.
Professional experience
Fraud, risk & complex operations
I progressed through billing, refund and complex operations before moving into Revenue Protection investigations. Across those roles, I worked with suspicious activity, failed automated checks, conflicting data and complex cases where the correct outcome was not immediately obvious — identifying patterns, assessing risk, reconciling evidence and making decisions within a regulated environment.
- Fraud investigation
- Risk assessment
- Anomaly detection
- Financial risk
- Evidence
- Regulated decision-making
Independent work
Products, tools & experiments
Outside work, I build because I get curious. Sometimes that means shipping a consumer app. Sometimes it means building a Spotify library auditor, experimenting with automated trading or creating an operational dashboard because I was tired of information living in six different places. They cover very different problems on purpose.
- Product building
- APIs
- Automation
- Rules systems
- Dashboards
- AI-assisted development
What carries across both is curiosity, structured thinking and a willingness to follow a problem further than I originally expected.
02 — What I bring
What I bring
Grouped by where I've actually used it — including the things I'm still developing.
Used professionally
Investigation & judgement
- Suspicious activity investigation
- Anomaly and pattern detection
- Risk assessment
- Financial exposure assessment
- Evidence handling and documentation
- Working across conflicting data sources
- Complex case ownership
- False-positive awareness
- Regulated decision-making
- Turning findings into action
Tools
- Advanced Excel
- Google Sheets
- Case-management systems
Used professionally
Financial & complex operations
- Refund and account review
- Failed automated-check review
- Account reconciliation
- Payment and direct-debit checks
- Data reconciliation
- Exception handling
- Dispute resolution
- Cross-functional coordination
- Third-party liaison
- Process improvement
Used in independent projects
Building & systems
- AI-assisted development
- Requirements and specification writing
- Rules-based system design
- Automation
- Operational dashboards
- API integrations
- Database-backed applications
- Scoring and decision logic
- Testing failure conditions
- Data validation
Tools
- Next.js
- Supabase / Postgres
- Tailwind
- Capacitor
- RevenueCat
- API
- OAuth / PKCE
Currently developing
Building independent fluency
Building fluency
- SQL
- BigQuery
- Data visualisation
- Looker / Tableau
Some of my projects already use SQL and SQL-backed systems through AI-assisted development. I'm continuing to build my ability to work with SQL independently.
03 — Selected work
Things I've built
Different problems. Different reasons for building them. Each one shows a different part of how I work.
MELAIRE Trust & Revenue Console
An explainable abuse-detection and investigation workflow for my own subscription app. Eight detection rules feed a review queue, weighted risk score and case evidence view, while resolved cases feed into per-rule precision so false positives are visible rather than hidden.
- Detection rules
- Risk scoring
- Postgres
- Next.js
MELAIRE
A hair-care platform for iOS. A guided profile, curated ingredient data, product analysis and personalised outputs — built around one rule I care about: if the system doesn't know, it shouldn't pretend that it does.
- Next.js
- Supabase
- Capacitor
- RevenueCat
Playlist Cleanup
I wondered how messy my Spotify library actually was. So I built a read-only auditor for duplicates, re-releases, suspiciously similar tracks, stale playlists and ghost tracks — complete with confidence scores and the evidence behind each finding.
- Spotify API
- OAuth PKCE
- Entity resolution
- Next.js
Algorithmic trading systems
Automated futures and FX systems built around explicit trading rules, hard loss limits, kill switches and simulation-before-live. Less “predict the market”, more “make absolutely sure the system knows when it is not allowed to act.”
- Rules engine
- Risk controls
- CME
- MT5
Let's talk.
I'm interested in difficult problems, unusual systems and work where asking the next question matters. If you think my background could be useful to your team, I'd be happy to talk.