Skip to content

Jordan Browne-Moore

I build and ship ML systems for fintechs — LLM training, credit risk models, and fraud pipelines — from prototype through regulated deployment.

Companies hire me to build these systems, then retain me to keep them running.

LLM Training & MLOps  ·  Credit Risk & Scoring  ·  Fraud & KYC Automation

50%
Reduction in first-payment defaults
42x Faster
KYC verification turnaround
80%
Cost reduction vs third-party LLM APIs
See My WorkDiscuss Your Project

Fully remote contracts  |  UK & US clients  |  GMT to GMT+8 availability

Where I've shipped production systemsKuda Technologies  ·  CB Insights  ·  Capgemini

Neobanks scaling their lending products. Fintechs automating compliance workflows. Market intelligence platforms replacing expensive API dependencies with custom trained models.

Work

50% Default Reduction — Credit Scoring System

Kuda Technologies (African neobank, 7M+ customers)

Problem

Needed to pre-qualify millions of customers monthly for lending products. High default rates were eroding portfolio performance.

What I Built

Designed and owned the credit scoring function: an optimised LightGBM model built from a 300K borrower sample, processing millions of monthly pre qualification decisions. Developed the feature stability methodology (bootstrap learning curves) that became the team's standard for all subsequent model builds.

Results
50%
reduction in first-payment defaults

50% reduction in first-payment defaults. 75% reduction in total portfolio defaults. Model serves as the production scoring engine for the entire lending portfolio.

LightGBMPythonSciPySHAPSQLAWSBootstrap framework

80% Cost Reduction — LLM Fine-Tuning for Revenue Estimation

CB Insights, Market Intelligence Platform

Problem

Needed to estimate private company revenue from public signals at scale. The initial solution relied on expensive third-party LLM API calls, making costs prohibitive at platform scale.

What I Built

End to end product build: designed the data pipeline, normalisation layer, fine tuning approach (Qwen 3 32B, LoRA), and production FastAPI inference API. Defined the evaluation methodology (format validity, near match accuracy, and ground truth proximity), establishing how the team measures model quality for all subsequent LLM work.

Results
80%
cost reduction vs third-party LLMs

80% reduction in API costs vs third-party LLMs. Production inference API serving revenue estimates across the platform.

Qwen 3 32BLoRAFastAPIPythonHuggingFace Transformers

7 Days to 4 Hours — KYC Automation Pipeline

Kuda Technologies

Problem

KYC verification took 7 days end to end, requiring a large manual review team. A hard bottleneck to customer onboarding at scale.

What I Built

Scoped, designed, and shipped an end to end KYC automation system using vision language models (Gemini Flash, Qwen3 8B VL), taking it from initial problem definition through model selection, fine tuning, and production deployment. System classifies identity documents, detects fraudulent submissions, and auto approves customers exceeding confidence thresholds.

Results
95%
reduction in manual review headcount

Verification time reduced from 7 days to 4 hours. Manual review headcount reduced by 95%. System now handles the majority of KYC decisions autonomously, with human review reserved for edge cases only.

Gemini FlashQwen3 8B VLPythonPrompt engineeringVLM fine tuning

More Work

Graph-Based Fraud Ring Detection

Kuda Technologies

Problem

Organised fraud rings operating across multiple connected accounts were evading individual-account detection methods.

What I Built

Initiated and built a graph-based fraud detection capability that did not previously exist at the company. NetworkX-based analysis mapping transaction relationships and behavioural patterns to surface organised multi-party fraud networks.

Results

Identified and disrupted 20+ fraud rings (3+ connected accounts each). Output fed directly into risk decisioning, triggering account restrictions and recovery actions across affected networks.

NetworkXPythonGraph analysisSQL

Salary & Recurring Inflow Detection

Kuda Technologies

Problem

Lending affordability assessment required reliable detection of salary and recurring income. Third-party supplier (Mono) had limited coverage, particularly on accounts with sparse transaction history.

What I Built

Designed an algorithmic framework for salary and recurring inflow detection from raw transaction data, replacing reliance on a third-party supplier with limited coverage. System directly feeds into credit eligibility and affordability decisioning.

Results

35% coverage uplift over the third-party supplier. 45% uplift on accounts where the supplier returned no data. 9% false positive rate. Directly improved credit eligibility accuracy for underbanked customers.

PythonSQLAlgorithmic design

Open Source

ElSnacko/llm-steering

LLM Activation Steering Toolkit

Python research toolkit for fine-grained control of LLM refusal behaviour. Domain-aware activation steering with per-category vectors, going beyond simple abliteration to independently target specific behaviour types.

Independent research into LLM internals: activation-level control, not API-level prompting.

ElSnacko/feature-bootstrapping-toolkit

Feature Bootstrapping Toolkit

Bootstrap learning curve framework for feature stability testing in production ML models. Determines how much data a feature needs before its predictive signal stabilises, whether the feature has abundant historical outcomes or no prior observations at all. Handles both standard supervised features and the harder case of sparse or absent observation data, which is common in emerging market lending, thin-file populations, and new product launches.

The methodology behind the 50% default reduction in the credit scoring case study, generalised and open-sourced.

ElSnacko/activation-steering-asymmetry

Activation Steering Asymmetry

Research toolkit investigating why steering LLMs toward compliance fails to unlock harmful outputs. Demonstrates that the refusal axis is layer-local and steering is asymmetric: pushing models toward refusal is cheap and effective, while pushing toward compliance degrades output before producing it. Includes per-category steering vectors, bootstrap stability analysis, and dual attention+MLP extraction across Qwen3.5-9B and Mistral-7B.

Reproduces the published claims of arXiv:2512.16602, with a full verification report.

Articles

Technical analysis and methodology from production ML, credit risk, and applied LLM work.

Background

Senior Data Scientist → Consultant — Kuda Technologies

Hired to build the credit risk and fraud detection functions from the ground up. Retained as external consultant to continue leading these workstreams after transitioning out of the full-time role. Owned model development, validation methodology, and production deployment across credit scoring, KYC automation, and fraud detection. Mentored 4 data scientists and participated in cross-seniority hiring.

Data Scientist III — CB Insights

LLM fine tuning, NLP, production ML engineering.

Senior Data Science Consultant — Capgemini

Financial services consulting.

Data Science Consultant — Beyond Analysis
MSc Financial Economics — Birkbeck, University of London

Dissertation: Predicted weekly equity movement with 80% accuracy using sentiment data and deep learning.

BSc Economics & Finance — Southern Oregon University

Certifications

AWS Cloud Practitioner  |  GCP Associate Cloud Engineer  |  Azure AZ-900

Get in Touch

Available for LLM, credit risk, fraud, and KYC engagements. Book a call to discuss your project or send an email if you prefer async.

Discuss Your Project
[email protected]github.com/ElSnackolinkedin.com/in/jordan-j-browne-moore

Available for fully remote contracts. UK, US, and international clients.