Niral J Shah
Say hello

FinTech / Data / Plain writing

I build things you can inspect.

Lending software. Measured language. Data projects with the working attached.

Explore the work Read the story
SOURCE RULE EVIDENCE SHOW THE WORKING. TRACEABLE OUTPUT
  • FUNDO LOANS$1M MRR → 3× multiple at acquisition, 1.5× the loan book
  • CREDITCREST23 lending products built
  • MUNDANE READ27 news feeds · 17 mastheads
  • HOSTED AI0 models called
  • PACKAGES0 third-party dependencies
  • AUTOMATIONover $265K saved, approved by PwC
  • OCRaccuracy 10% → 74%
  • REPORTINGdata-to-report 14× faster
  • MANUFACTURINGUS$38,150 saved in 2 months
  • TOOLSfault prediction 65% → 90%
  • LOGISTICS74.45% accurate, built in 15 days
  • TAGGERadjective F1 0.80 on news
  • PYPIpip install basicanalysis
  • WALKING CHALLENGE10.5M steps in 8 weeks

Selected work

Two products. One principle.

Creditcrest Technologies

23 built & testedPre-launch

Lending infrastructure for Australian credit providers: twenty-three products, built and tested, from the first screen of an application to the last report to the credit bureau. Each can be used on its own, and every figure comes with the evidence behind it.

Try every product (opens in a new tab) · Creditcrest Technologies is not a credit provider, is not a credit assistance provider, and does not hold an Australian credit licence. Its software produces evidence; the licensee makes the decision.

Explore Creditcrest (opens in a new tab)

Measured, not claimed

One story, three versions

Quest apartments data breach · Niral Score, lower is plainer

  1. 7NEWS9.9
  2. ABC News7.8
  3. Mundane Read7.6
As published on mundaneread.com, September 2026.

Hi, I’m Niral

I turn messy financial and language data into numbers you can check.

I trained as an electronics engineer and have spent the decade since where data meets money. As lead data scientist at Fundo Loans, I helped a small team take an Australian lender from $1M MRR to a 3× multiple at acquisition, with 1.5× the loan book, building its data architecture, automated decisioning and bank-statement analysis from scratch.

Before that: credit and fraud models for lenders, insurance risk at PD Insurance, yield engineering at Maxim Integrated (Now Analog Devices), and a master’s in Data Science at Monash. On the side, I’m building software for lending and for the news, with one rule for both: every number comes with the working that produced it.

Off the clock: table tennis, stargazing and listening to big band.

Worked withFundo LoansCash DirectTrimbleMonash UniversityPD InsuranceMaxim Integrated (Now Analog Devices)

The full story & CV
Building Creditcrest & Mundane Read · Parramatta, NSW

Back in school, my teacher docked marks on my report writing. She explained that we should not use removable adjectives, overly descriptive words, or include personal opinions.

If only she had seen the news lately.

From electronics to evidence.

  1. 2011–15 B.E. Electronics & Communication Vishwakarma Government Engineering College
  2. 2016–18 Yield Engineer Maxim Integrated (Now Analog Devices)
  3. 2018–19 Data Analyst PD Insurance
  4. 2019–20 Research Assistant Monash University, SoDa Labs
  5. 2019–21 Data Scientist, self-employed Dusk Mobile · Trimble · Cash Direct
  6. 2020 Master of Data Science Monash University
  7. 2021–now Lead Data Scientist Fundo Loans
  8. Now Building on the side Creditcrest Technologies · Mundane Read

Now

What I’m building next.

Updated

Creditcrest

  1. IntakeTransactions by consent under the Consumer Data Right. Waits on accreditation or a sponsor.
  2. A hosted ApplyAn application a lender can brand, with identity, address prefill and statements in one flow.
  3. Funder reportingWarehouse borrowing-base and tape reports on a schedule, from the LoanManager book.

Mundane Read

  1. A framing benchmarkNews photographs described by working picture editors, with the scores compared against theirs.
  2. An accuracy benchmarkA held-out set scored by working journalists, with the agreement rate published whether or not it flatters.
  3. AppsiOS and Android, then macOS and iPadOS.
  4. An editorial linterThe neutralizer as a command-line tool and a continuous-integration check for copy desks.

Projects

Things I’ve built, measured and shipped.

All 11 projects

  • Now / Live

    Adjective tagger, trained in-house (opens in a new tab)

    A part-of-speech tagger trained on annotated English. On news it finds adjectives at F1 0.80, where matching word endings scores 0.56 on the same text.

    The Method page

  • 2019

    Stock price analyser

    Daily sentiment of news about a company read against its share price. The ancestor of Mundane Read.

    Read the write-up

  • 2019

    ELO-rating model for loan default

    A credit score built the way chess ranks players, from bank statements alone, with a confidence interval that says when to buy a credit report instead.

    Read the write-up

  • 2019

    basicanalysis (opens in a new tab)

    A Python package, under a hundred lines, that runs the core metrics of most supervised-learning methods over a dataset in one go.

    pip install basicanalysis

  • 2019

    Fall Relief

    A team of four building targeted exercise and diet plans to reduce falls among people aged 65 and over.

    Read the write-up

  • Next

    The next one is in the works.

    I’m always building something. New projects land here first, and in the newsletter.

    Hear about it first

What I know well

FinTech, data, and the words wrapped around both.

  • Lending & credit

    Automated decisioning and bank-statement analysis at Fundo Loans. On the side, twenty-three lending products at Creditcrest, each showing the evidence behind every number.

    • Decisioning
    • Serviceability
    • Consumer Data Right
    • Best interests duty

  • Data science & ML

    Credit scores and fraud models for lenders, an ELO-style credit score, a part-of-speech tagger trained on annotated English: each measured against the question it is meant to answer.

    • Supervised learning
    • NLP
    • Sentiment
    • Forecasting

  • Data pipelines

    Pipelines reading 27 news feeds around the clock, each on its own schedule, scoring every article on seven indicators and keeping the evidence. No third-party packages.

    • Ingestion
    • D1 / SQLite
    • Workers
    • Zero dependencies

  • Measuring language

    Sentiment, framing and sensational wording, counted with the words behind each number. The Niral Score folds seven readings into one figure from 0 to 100.

    • Niral Score
    • Evidence spans
    • Explainability

Writing

From the archive.

Write-ups from before Creditcrest. New writing starts soon, and the newsletter gets it first.

All writing

/ FinTech / 1 min

An ELO-rating model for loan default prediction

Scoring loan applicants the way chess rates players — against a model applicant who repaid — with a confidence interval that says when to buy a credit report.

Read the article

Signal over Noise

A build. A number. A good read.

Occasional notes on FinTech, data and plain writing: one thing I built, one number worth knowing with its source, and one thing worth reading. Sent when there is something worth sending.

Free and occasional. Email confirmation first, one click to unsubscribe, and no tracking pixels. Privacy

Contact

A useful conversation starts here.

Lending software, data work, Mundane Read, or something you read here. Leave your email and I’ll reply to you personally.

Your name, email and message are kept so I can reply, and nothing else is done with them. Privacy